Authors: Natalia Anna Koc, Maurycy Rakowski, Samuel D. Pettersson, Adriana Mika, Piotr Zieliński, Tomasz Szmuda
Categories: Review, Oxidative stress, Biomarkers, Subarachnoid hemorrhage, Cerebral vasospasm, Delayed cerebral ischemia
Source: Neurosurgical Review
Authors: Natalia Anna Koc, Maurycy Rakowski, Samuel D. Pettersson, Adriana Mika, Piotr Zieliński, Tomasz Szmuda
Intracranial aneurysms often remain asymptomatic until rupture, causing aneurysmal subarachnoid hemorrhage (aSAH). aSAH frequently leads to cerebral vasospasm (CVS) and delayed cerebral ischemia (DCI), significantly increasing the risk of severe neurological deficits and mortality. Identifying reliable biomarkers, such as lipid peroxidation metabolites (LPMs), is crucial for early prediction and timely intervention. This study summarizes current knowledge on LPMs as potential biomarkers for CVS and DCI after aSAH. A systematic review was conducted following PRISMA guidelines. Two independent authors searched PubMed, Web of Science, and Scopus for articles studying the association between non-enzymatic and enzymatic lipid metabolites and CVS or DCI after aSAH. Quality and risk of bias were evaluated using the Newcastle-Ottawa Scale. Extracted data included metabolite concentrations, biological sample types, timing of collection, patient demographics, clinical severity of aSAH, Fisher’s grade, DCI definition, and relationship to DCI. Of 519 records screened, 17 studies were included. Lipid metabolites were measured in blood (5 studies), cerebrospinal fluid (11 studies), and urine (2 studies). F2-isoprostanes (F2-IsoPs), studied in 7 articles, were linked to increased DCI risk, with elevated levels observed within three days post-aSAH. Isofurans (IsoFs) predicted DCI risk between days 5 and 8 post-aSAH, while elevated cholesteryl ester hydroperoxide (CEOOH) levels on day 2 linked to symptomatic vasospasm. Enzymatic arachidonic acid (AA) metabolites, including 6-keto-prostaglandin F1-α, prostaglandin D2, and leukotriene C4, were also associated with early DCI risk. To the best of our knowledge, this review is the first to comprehensively assess all LPMs in relation to CVS and DCI. Elevated concentrations of F2-IsoPs and enzymatic AA derivatives may serve as biomarkers for DCI prediction in aSAH. These findings highlight the need to explore the potential of LPMs, paving the way for risk stratification and timely interventions to improve patient outcomes and aid researchers in developing predictive scoring systems for DCI.
Clinical trial number Not applicable.
The online version contains supplementary material available at 10.1007/s10143-025-03662-3.
The worldwide prevalence of intracranial aneurysms (IAs) is estimated at 3.2%, and they often remain asymptomatic for extended periods, making them a latent risk to neurological health [19, 58]. Early detection and intervention are critical to reduce the risk of rupture, which can lead to aneurysmal subarachnoid hemorrhage (aSAH) - a serious condition that results in cerebral vasospasm (CVS) within two weeks of the initial hemorrhage in approximately 67% of cases [10]. CVS manifests as localized or generalized contraction of the cerebral arteries and clinically presents as a new focal neurological deficit or a reduction in Glasgow Coma Scale (GCS) score after excluding other causes such as rebleeding or hydrocephalus [12].
CVS can progress to delayed cerebral ischemia (DCI), a serious complication where arterial constriction limits blood flow and leads to oxygen deprivation in brain tissue [12]. DCI increases the risk of cerebral infarction, disability, and mortality, causing symptoms such as focal neurological deficits, reduced consciousness, and confusion [48, 50]. Even if not fatal, DCI can result in permanent functional impairment and reduced quality of life.
Evidence suggests that the production of reactive oxygen species (ROS) is crucial for the occurrence of DCI [12]. The brain’s high susceptibility to oxidative stress (OS) comes from its high oxygen consumption and unique metabolic processes, such as neurotransmitter oxidation [38]. This susceptibility is heightened by the brain’s relatively modest antioxidant defenses and high content of polyunsaturated fatty acids (PUFAs) in cellular membranes, especially arachidonic acid (AA, 4n-6) and docosahexaenoic acid (DHA, 6n-3), which are particularly ROS-sensitive [21, 35]. Excessive ROS disrupt cellular redox balance, leading to cell dysfunction or death [54]. Increased free radical formation after aSAH, combined with an inadequate endogenous antioxidant response, can exacerbate the damaging effect.
ROS-induced OS can affect various biological pathways, as shown by the results of Klepinowski et al., who investigated possible epigenetic mechanisms underlying CVS and DCI [24]. Biochemically, peroxidation by free-radicals involves ROS reacting with PUFAs in cell membranes, forming stable prostaglandin-like compounds (PLCs) that act as OS biomarkers. Non-cyclooxygenase peroxidation products of AA are particularly reliable indicators of lipid peroxidation in vivo [34]. Beyond indicating oxidative damage, PLCs contribute to cerebral vasoconstriction and platelet activation, linking lipid peroxidation metabolites (LPMs) to vascular pathologies [30].
Lipid metabolites produced via cyclooxygenase (COX) and lipoxygenase (LOX) pathways also play a critical role in OS. These enzymatic processes involve peroxidation or hydroperoxidation steps, generating various bioactive end-products, including prostaglandins (PGDs), leukotrienes (LTCs), and thromboxanes (TBXs), that serve as markers and mediators in OS and inflammation [31]. Their production significantly increases after tissue injury, fueling the inflammatory cascade and influencing vascular tone [45].
To our knowledge, no systematic review has examined the role of enzymatic and non-enzymatic PUFA metabolites, also called oxylipins, in CVS or DCI after aSAH. Given the global burden of these conditions and the lack of effective prevention or treatment, identifying new diagnostic biomarkers and pharmacological targets is crucial. This study summarizes current knowledge on LPMs as potential biomarkers for CVS and DCI in aSAH patients.
The screening process was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The PubMed, Web of Science, and Scopus databases were used to screen studies from inception to July 30, 2024, using following (lipid OR lipoprotein OR cholesterol) AND (peroxid* OR oxid* OR arachidonic acid OR thromboxane OR leukotriene* OR prostacycline* OR isoprostane* OR polyunsaturated fatty acids) AND (intracranial aneurysm OR aneurysmal) AND (SAH OR subarachnoid hemorrhage OR cerebral vasospasm OR cerebral ischemia). Two authors (N.A.K. and M.R.) independently conducted the database search. The study protocol was registered via PROSPERO (registration no. CRD42024549985, https://www.crd.york.ac.uk/PROSPERO/).
The inclusion criteria for the studies were reporting patients with aSAH, concentrations of enzymatic or non-enzymatic lipid metabolites measured in biological specimens, sample size of at least 10 patients, and being written in English. Exclusion criteria animal studies, extracranial aneurysms, lack of a control group of non-aSAH or non-CVS/DCI patients. Extractable data was required for inclusion. Two authors (N.A.K. and M.R.) independently screened the article titles and abstracts using Rayyan (https://rayyan.qcri.org/). The studies that passed the initial screening phase were reassessed in a full-text screening phase.
The data of interest included LPMs concentrations in patients with aSAH and CVS or DCI. The following variables were also biological specimen type, timing of sample collection, outcome (aSAH, IAs, CVS or DCI), number of patients, gender, age, clinical severity, Fisher’s grade, DCI definition, significant metabolite concentration and its timing, relationship to DCI. Early metabolite concentrations were defined as those measured at the first time point and if unavailable, mean concentrations were reported. Lumbar CSF concentrations were provided for CSF-based measurements. Data were organized into tables, and predictive metabolite concentrations for DCI cases were analyzed using mean difference (MD) with 95% confidence intervals (CI) via random-effects models. Heterogeneity was assessed using the I² statistics. Publication bias was assessed by visual inspection of funnel plot asymmetry. Statistical significance was set at p < 0.05. Analysis was performed in Revman (Version: 8.17.0).
The quality assessment of each included study was conducted using the Newcastle-Ottawa Scale (NOS). A maximum of 2 points was awarded for the comparability category, and up to 1 point for each remaining domain. A total score of ≥ 7 indicated high quality, 4–6 moderate quality, and ≤ 3 low quality. The assessment was performed by one author (M.R.).
The literature search identified 519 articles, with 488 excluded after title and abstract screening. The remaining 32 studies underwent full-text analysis to assess adherence to predefined criteria, resulting in including 17 articles. The PRISMA diagram, detailing the literature selection process, is presented in (Fig. 1). Quality assessment showed sixteen studies as high quality and one as moderate quality (Supplementary Table 1). Target metabolites were categorized by lipid groups, including free fatty acids (FFAs), eicosanoids, PLCs, and reactive aldehydes. Study characteristics were organized by biological specimen type in Tables 1, 2 and 3. Metabolite measurements were performed using blood in five studies (Table 1), CSF in eleven studies (Table 2), and urine in two studies (Table 3). A comparison between DCI and non-DCI patients, focusing on identified LPMs and their clinical significance, is summarized in (Table 4). One article studied F2-isoprostanes (F2-IsoPs) and F4-neuroprostanes (F4-NPs) concentrations in IAs and non-IAs patients, with details presented in (Table 5). Summary of MD analysis is presented in (Table 6) and forest plots are included as Supplementary Fig. 1–4. The funnel plot showed asymmetry in study distribution, suggesting publication bias. (Figure 2). shows an acyclic graph for aSAH, PUFA peroxidation processes, and DCI.
Fig. 1PRISMA diagram showing selection process of the studies. Data added to the Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097. 10.1371/journal.pmed1000097
Table 1Characteristics of the study investigating metabolites in bloodStudyCountryLipid groupTarget metaboliteGroups of patientsTime from aSAH to sample collectionObserved outcomeNo. of patientsSubjects with DCI/CVSEarly target metabolite concentration* (DCI vs. no-DCI)Early target metabolite concentration* (SAH vs. no-SAH)Females, n (%)Mean age (SD)Hunt Hess scale, nSAH Fisher’s grade, nSyta-Krzyżanowska 2018 [56]PolandProstaglandin-like CompoundsF2-IsoPsIA vs. non-IA, SAH vs. non-SAHTime points at 1st, 3rd, 6-8th dayaSAH335449 ± 1086 vs. 1192 ± 575 pg/dL1079 ± 654 vs. 442 ± 119 pg/dL35 (53)50.8 (NR)1, n = 191–2, n = 11F4-NPs336854 ± 8596 vs. 5562 ± 2446 pg/dL6220 ± 3869 vs. 660 ± 194 pg/dL2–4, n = 143–4, n = 22Kaynar 2005 [23]TurkeyReactive AldehydesMDASAH vs. non-SAHTime points at 1st, 3rd, 5th, and 7th day from aSAH onsetaSAH217NR3.0 ± 0.4 vs. 2.4 ± 0.5 umol/mL13 (62)46.4 (15)1, n = 01, n = 02, n = 132, n = 113, n = 53, n = 94, n = 24, n = 15, n = 1Seifert 1987 [51]GermanyEicosanoids6-keto-PGF1αSAH vs. non-SAHWithin 1 to 5 days from aSAH onsetaSAH124NR113.0 (20–542) pg/mL vs. below detection level (< 20 pg/mL)10 (83)50.8 (12)1, n = 4NRTXB2NR56.9 (20–143) pg/mL vs. NR2, n = 63, n = 2Wisniewski 2023 [61]PolandProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAH, DCI vs. non-DCITime points at 2nd, 4th, and 6th day after aSAH onsetaSAH, DCI451764.1 ± 31.6 vs. 40.6 ± 21.3 pg/mL49.5 ± 27.8 vs. 21.1 ± 5 pg/mL26 (58)NR1, n = 21, n = 32, n = 152, n = 173, n = 253, n = 164, n = 34, n = 9Lin 2006 [33]TaiwanProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAHNRaSAH153NRF2-IsoPs: 17.6 ± 3.4 vs. 12.0 ± 1.2 pg/mL9 (60)62.5 (15.3)1, n = 11, n = 22, n = 82, n = 23, n = 33, n = 114, n = 34, n = 6* Early metabolite concentration was defined as the one measured at the first time pointF2-IsoPs: F2-Isoprostanes; F4-NPs: F4-Neuroprostanes; MDA: Malondialdehyde; 6-keto-PGF1α: 6-keto-prostaglandin F1-α; TXB2: Thromboxane B2
Table 2Characteristics of the studies investigating metabolites in CSFStudyCountryLipid groupTarget metaboliteGroups of patientsTime from aSAH to sample collectionObserved outcomeNo. of patientsSubjects with DCI/CVSEarly target metabolite concentration* (DCI vs. no-DCI)Early target metabolite concentration* (SAH vs. no-SAH)Females (%)Mean age (SD)Hunt Hess scale (n, %)SAH Fisher’s grade, n (%)Kamezaki 2002 [22]JapanReactive AldehydesPCOOHCVS vs. non-CVSWithin 48 h from aSAH onset, and 6th or 7th dayaSAH, CVS20130.25 ± 0.4 nmol/mL vs. NRNR14 (70)55.5 (12.9)1, n = 11, n = 1CEOOH0.19 ± 0.34 nmol/mL vs. NR2, n = 82, n = 03, n = 73, n = 144, n = 44, n = 5Seifert 1987 [51]GermanyEicosanoids6-keto-PGF1αSAH vs. non-SAHWithin 1 to 5 days from aSAH onsetaSAH124NR113.0 (20–542) pg/mL vs. below detection level < 20 pg/mL10 (83)50.8 (12)1, n = 4NRTXB21575 (26-9382) pg/mL vs. below detection level < 20 pg/mL2, n = 63, n = 2Pilitsis 2002 [46]USAFree Fatty AcidsMyristic acidSAH vs. non-SAHTime points at 1st day, and between 8th-10th day from SAH onsetSAH205NR403 ± 83 vs. 160 ± 16 ug/mL12 (60)50 (31–75)1, n = 31, n = 2Docosahexanoic acid365 ± 70 vs. 61 ± 6 ug/mL2, n = 92, n = 4Arachidonic acid151 ± 34 vs. 26 ± 3 ug/mL3, n = 43, n = 8Linoleic acid261 ± 68 vs. 105 ± 11 ug/mL4, n = 44, n = 6Palmitic acid993 ± 118 vs. 638 ± 37 ug/mLOleic acid420 ± 73 vs. 128 ± 10 ug/LGaetani 1986 [14]ItalyEicosanoids6-keto-PGF1αSAH vs. non-SAH, CVS vs. non-CVSFirst sample within 1 to 4 days from aSAH, next samples at different time points between 2nd and 14th day after aSAH onsetSAH, CVS123399.3 ± 103.6 vs. 184 ± 122.7 pg/mL237.83 ± 149.6 vs. < 10 pg/mL6 (50)51.3 (16)1, n = 2NR2, n = 73, n = 3Rodriguez y Baena 1988 [44]ItalyEicosanoidsPGD2SAH vs. non-SAH, CVS vs. non-CVSWithin 1 to 3 days from aSAH onsetaSAH, CVS40161129.62 ± 146.33 vs. 460.10 ± 55.89 pg/mL367.50 ± 47.42 vs. 74.10 ± 17.3 pg/mLNRNR1, n = 19Kistler 1, n = 186-keto-PGF1α6214.20 ± 19.96 vs. 306.37 ± 88.74 pg/mL6236.10 ± 30.89 vs. 21.90 ± 6.6 pg/mL2, n = 142, n = 15LTC42582.19 ± 381.83 vs. 812.92 ± 142.06 pg/mL335.50 ± 38.81 vs. 107.50 ± 42. pg/mL3, n = 73, n = 7TXB24350.25 ± 656.87 vs. 5752.50 ± 899.25 pg/mL1009.72 ± 101.87 vs. 267.16 ± 41.6 pg/mLRodriguez y Baena 1987 [49]ItalyEicosanoidsPGD2SAH vs. non-SAHWithin 1st and 3rd day from aSAH onsetaSAH, CVS3012615 ± 105.5 vs. 208.3 ± 38.9 pg/mL371 ± 328.17 vs. 79.0 ± 17.3 pg/mL18 (43)47.1 (11.2)1, n = 81, n = 156-keto-PGF1α6260.7 ± 48.4 vs. 213.6 ± 57.9 pg/mL249.04 ± 213.91 vs. 21.9 ± 6.6 pg/mL2, n = 172, n = 53, n = 53, n = 10Asaeda 2005 [1]JapanProstaglandin-like Compounds8-iso-PGF2α = F2-IsoPsSAH vs. non-SAHWithin 48 h, between 3-5th day, 6-8th day, 9-11th day, 12-14th day, and 20th day after the aSAH onsetaSAH, CVS341942.4 ± 37.1 vs. 24.8 ± 12.0 pg/mLNR22 (65)62.7 (41–91)NR2, n = 43, n = 294, n = 1Hsieh 2009 [18]TaiwanProstaglandin-like CompoundsF4-NPsSAH vs. non-SAHNRaSAH153NR250.6 ± 244.4 vs. 77.8 ± 19.5 pg/ml10 (67)64.2 (15)1, n = 11, n = 22, n = 82, n = 23, n = 33, n = 114, n = 34, n = 6Lin 2006 [33]TaiwanProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAHNRaSAH153NR27.2 ± 5.7 vs. 8.7 ± 1.2 pg/mL9 (60)62.5 (15.3)1, n = 11, n = 22, n = 82, n = 23, n = 33, n = 114, n = 34, n = 6Gomes 2022 [16]USAProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAH, DCI vs. non-DCIWithin 24 h hours from aSAH onset, and between 5-8th day after aSAHaSAH, DCI18741 (33.5–52) vs. 44 (28.5–55.5) pg/mL37 ± 11 vs. 28.17 ± 8.04 pg/mL14 (78)61.2 (15.7)3 (3–4)^a^3 (3–4)^a^IsoFs57 (34–72) vs. 0 (0–34) pg/mL4.67 ± 14 pg/mL vs. below detection levelWisniewski 2024 [63]PolandProstaglandin-like Compounds8-iso-PGF2α = F2-IsoPsDCI vs. non-DCI1st, 3rd, and 5th day post-surgery (surgery within 24 h after aSAH)DCI271365.7 ± 50.67 vs. 23.7 ± 19.20 pg/mLNR11 (41)60 (50–66)1, n = 3Modified Fisher’s 1, n = 32, n = 82, n = 43, n = 173, n = 44, n = 14, n = 16* Early metabolite concentration was defined as the one measured at the first time point^a^ Mean value6-keto-PGF1α: 6-keto-prostaglandin F1-α; TXB2: Thromboxane B2; PGD2: Prostaglandin D2; LTC4: Leukotriene C4; 8-iso-PGF2α: 8-iso-prostaglandin F-2α; Isofurans: Isofurans
Table 3Characteristics of the studies investigating metabolites in urineStudyCountryLipid groupTarget metaboliteGroups of patientsTime from aSAH to sample collectionObserved outcomeNo. of patientsSubjects with DCI/CVSEarly target metabolite concentration* (DCI vs. no-DCI)Early target metabolite concentration* (SAH vs. no-SAH)Females (%)Mean age (SD)Hunt Hess scale (n, %)SAH Fisher’s grade, n (%)Wiśniewski 2017 [60]PolandProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAH, DCI vs. non-DCIEvery day between 2nd and 6th day from aSAH presentationaSAH, DCI20918.21 ± 7.34 vs. 16.07 ± 7.02 pg/1 mg creatinine17.02 ± 7.03 vs. 6.9 ± 0.9 pg/1 mg creatinine11 (55)60 (13.3)1, n = 41, n = 32, n = 62, n = 43, n = 83, n = 24, n = 24, n = 11Wiśniewski 2020 [62]PolandProstaglandin-like CompoundsF2-IsoPsSAH vs. non-SAH, DCI vs. non-DCIEvery day between 1st and 5th day after aSAH presentationaSAH, DCI381821.16 ± 18.64 vs. 16.27 ± 7.99 pg/1 mg creatinine18.65 ± 14.19 vs. 7.1 ± 0.9 pg/1 mg creatinine24 (63)59.2 (15.5)1, n = 41, n = 32, n = 112, n = 93, n = 233, n = 104, n = 16* Early metabolite concentration was defined as the one measured at the first time pointF2-IsoPs: F2-Isoprostanes
Table 4Characteristics of the studies investigating predictive values of LPMs in DCI patientsStudyTarget MetaboliteSpecimenDCI DefinitionNo. of DCI patients% of aSAH patientsMetabolite Concentration (DCI vs. non-DCI)Metabolite Concentration TimingRelationship to DCISyta-Krzyżanowska 2018 [56]F2-IsoPsBloodNR515Not extractable3rd day after aSAHNo relationF4-NPs68.54 ± 85.96 vs. 55.62 ± 24.46 pg/mL1st day after aSAHHigher concentration with a slower decrease than in controls.Wiśniewski 2023 [61]F2-IsoPsBloodDeterioration in a patient’s condition characterized by an unexplained reduction in the GCS score by ≥ 2 points, with or without focal neurological deficits lasting more than 1 h, after excluding other potential causes of neurological decline. Diagnostic confirmation included TCD showing MCA flow velocity exceeding 120 cm/sec and a Lindegaard ratio (MCA/ICA flow velocity) greater than 3. DSA was performed to confirm arterial narrowing.173864.1 ± 31.6 vs. 40.6 ± 21.3] pg/mL2nd day after aSAHPlasma F2-IsoPs concentrations on day 2 predicted DCI (AUC 0.733, 95% CI 0.575–0.891, p < 0.01).Kamezaki 2002 [22]PCOOHCSF1) Clinical: Angiographically observed reduction in vessel diameter greater than 50%, accompanied by contralateral motor weakness or global neurological deterioration occurring after day 3, indicative of anterior diffuse vasospasm. 2) Angiographic: A reduction in baseline vessel diameter by 25% on angiogram, which may or may not be associated with clinical deterioration.13650.25 ± 0.4 nmol/mL vs. NRWithin 48 h after aSAHNo relationCEOOH0.19 ± 0.34 nmol/mL vs. NRWithin 48 h after aSAHTendency for higher levels of CEOOH associated with the occurrence of symptomatic vasospasm (p = 0.002).Gaetani 1986 [14]6-keto-PGF1αCSF1) Severe: Mean arterial narrowing greater than 50%, combined with segmental reduction in diameter to less than 2 mm in the ICA or 1 mm in the A1 or M1 segments.2) Moderate: Mean arterial narrowing greater than 50% without the significant segmental narrowing observed in severe vasospasm.3) Mild: Mean arterial narrowing ranging from 20–50%.325399.3 ± 103.6 vs. 184 ± 122.7 pg/mLBetween 1st and 3rd day after aSAHNot statistically significant, but the levels of AA metabolites in cisternal CSF are proposed as biochemical evidence for the anatomical and radiological patterns of clotted blood, important for predicting vasospasm.Rodriguez y Baena 1988 [44]6-keto-PGF1αCSF1) Clinical transient or persistent neurological deficits with impaired level of consciousness and without CT evidence of rebleeding. 2) Angiographic Fisher’s 3 grade classification.1640214.20 ± 19.96 vs. 306.37 ± 88.74 pg/mLBetween 1st and 3rd day after aSAHNo relationTXB24350.25 ± 656.87 vs. 5752.50 ± 899.25 pg/mLBetween 1st and 3rd day after aSAHNo relationLTC42582.19 ± 381.83 vs. 812.92 ± 142.06 pg/mLBetween 1st and 3rd day after aSAHStatistical analysis reveals significantly elevated cisternal levels of LTC4 (P < 0.005) within 3 days of aSAH in patients exhibiting vasospasm.PGD21129.62 ± 146.33 vs. 460.10 ± 55.89 pg/mLBetween 1st and 3rd day after aSAHStatistical analysis reveals significantly elevated cisternal levels of PGD2 (P < 0.005) within 3 days of aSAH in patients exhibiting vasospasm.Rodriguez y Baena 1987 [49]PGD2CSF1) Clinical onset of transient or persistent neurological deficit with impaired level of consciousness and without CT evidence of rebleeding. 2) Angiographic Fisher’s 3 grade classification.1240615 ± 105.5 vs. 208.3 ± 38.9 pg/mLBetween 1st and 3rd day after aSAHMonitoring AA metabolites in the CSF following SAH could become a reliable indicator for predicting vasospasm risk.6-keto-PGF1α260.7 ± 48.4 vs. 213.6 ± 57.9 pg/mLBetween 1st and 3rd day after aSAHAsaeda 2005 [1]8-iso-PGF2α = F2-IsoPsCSFNR195686.2 ± 70.2 vs. 25.9 ± 17.4* pg/mLBetween 6th and 8th day after aSAHThe levels of 8-iso-PGF2 in the CSF of SAH patients with vasospasm increased in the days following the hemorrhage, peaking on days 6–8, which aligns with the period when cerebral vasospasm is most commonly observed (p < 0.05).Gomes 2022 [16]F2-IsoPsCSFFocal neurologic impairment or unexplained reduction in GCS of > 2 points lasting at least 1 h, not apparent immediately following aneurysm occlusion and not explained by other conditions.73941 ± 4.63 vs. 44 ± 6.75 pg/mLBetween 5th and 8th day after aSAHNo relationIsoFs57 ± 28.15 vs. 0 ± 25.19 pg/mLBetween 5th and 8th day after aSAHThe presence of IsoFs in CSF had 86% sensitivity to detect DCI, with a negative predictive value of 89%.Wiśniewski 2024 [63]8-iso-PGF2α = F2-IsoPsCSFUnexplainable new symptoms of confusion or drop in the level of consciousness (by ≥ 2 point in GCS), with or without accompanying focal neurologic deficits and confirmed by TCD or DSA.134865.7 ± 50.76 vs. 23.7 ± 19.20 pg/mL1st day after surgery, 2nd day after aSAHConcentration of IsoPs in CSF on the first day after surgery for a ruptured IA can serve as prognostic factors in DCI (AUC 0.791, 0.619–0.963).Wiśniewski 2017 [60]F2-IsoPsUrineClinical: Presenting symptoms of confusion or decreased level of consciousness with or without motor or speech neurological deficits. Angiographic: arterial narrowing or increase of blood flow velocity of over 120 cm/s in middle cerebral artery measured by TCD.94517.63 ± 5.22 vs. 11.75 ± 3.46 pg/1 mg creatinine3rd day after surgery, 4th day after aSAHStatistically significant association of higher F2-IsoPs levels on day 3 with higher risk of CVS (AUC 0.819, p < 0.05) and poorer long-term outcome after 1 to 4 months post-event (AUC 0.792, p < 0.05).Wiśniewski 2020 [62]F2-IsoPsUrineClinical: patients with neurological deterioration (confusion or decreased level of consciousness by at least 1 point on the GCS, with or without focal neurologic deficits, lasting for at least 1 h) after excluding other causes of neurologic deficits. Angiographic: Flow velocity in MCA exceeded 120 cm/s and the Lindegaard ratio (MCA low velocity/ICA flow velocity) greater than 3 in TCD and confirmation by the DSA.184718 ± 5.9 vs. 12.7 ± 5.1 pg/1 mg creatinine3rd day after aSAHStatistically significant association of higher F2-IsoPs levels on day 3 with higher risk of CVS (AUC 0.764, 95% CI: 0.606–0.922; p = 0.001, optimal threshold 13.8,sensitivity—0.778, specificity—0.800) and poorer outcome after 1- and 12-months post-event (AUC 0.685,95% CI: 0.507–0.863; p = 0.042 optimal threshold 8.8, sensitivity—0.471, specificity—0.952).* Mean concentration for all daysF2-IsoPs: F2-isoprostanes; F4-NPs: F4-Neuroprostanes; 6-keto-PGF1α: 6-keto-prostaglandin F1-α; TXB2: Thromboxane B2; LTC4: Leukotriene C4; PGD2: Prostoglandin D2; 8-iso-PGF2α: 8-iso-prostaglandin F-2α; IsoFs: Isofurans; DCI: delayed cerebral ischemia; GCS: Glasgow Coma Scale; CT: computed tomography; TCD: Transcranial Doppler; DSA: Digital Subtraction Angiography; MCA: middle cerebral artery; ICA: internal carotid artery
Table 5Characteristics of the study reporting LPMs concentration in patients with unruptured IAsStudyTarget MetaboliteSpecimenNo. of UIAs patientsEarly target metabolite concentration (IA vs. non-IA)Relationship to UIAsSyta-Krzyzanowska 2018 [56]F2-IsoPsBlood33991 ± 617 vs. 470 ± 126 pg/dLPresence of an aneurysm leads to an increase in the cyclized products of PUFA peroxidation (F2 isoprostanes and F4-neuroprostanes)F4-NPs5758 ± 2439 vs. 694 ± 215 pg/dLF2-IsoPs: F2-Isoprostanes; F4-NPs: F4-Neuroprostanes; UIAs: unruptured intracranial aneurysms
Table 6Results of MD analysis for concentrations predictive for CVS and DCILipid MetaboliteNo. of Reporting StudiesSpecimenMD (SE) [pg/mL]95% CIHeterogeneity: I^2F4-NPs1Blood12.92 (38.72)––F2-IsoPs1Blood23.50 (8.66)––PCOOH1CSFNA––CEOOH1CSFNA––6-keto-PGF1α3CSF39.83-90.86, 170.5195%TXB21CSF-1402.25 (246.29)––LTC41CSF1769.27 (99.76)––PGD22CSF537.24*279.69, 794.7996%F2-IsoPs2CSF-1.80-6.96, 3.3782%IsoFs1CSF57 (13.07)––F2-IsoPs2Urine5.58^a^2.27, 8.880%^a^ [pg/1 mg creatinine] statistically significant
Fig. 2Conceptual diagram showing the biochemical mechanism from aneurysmal subarachnoid hemorrhage to delayed cerebral ischemia. Created with BioRender.com
The brain’s lipid composition is crucial in aSAH and its complications, such as CVS and DCI. Brain tissue is rich in phospholipids and FAs, making it highly susceptible to OS and lipid peroxidation [42]. Given this structural predisposition, we conducted a comprehensive review of lipid disturbances in aSAH, CVS, and DCI, covering both non-enzymatic and enzymatic LPMs, as well as FFAs that act as substrates in various oxidation pathways. Baseline metabolite concentrations in aSAH patients without CVS or DCI were also analyzed to establish a comparative framework for lipid changes. We focused primarily on documenting early concentrations of the metabolites to explore their early prognostic value for identifying aSAH patients at risk of CVS and DCI.
The pathophysiology of DCI remains incompletely understood. It is influenced by extravasated blood volume in the subarachnoid space and associated with risk factors such as female sex, higher Hunt-Hess scale scores (≥ 4), modified Fisher grades (≥ 3), and pre-existing hypertension [64]. Among the proposed mechanisms, the presence of subarachnoid hemoglobin triggers ROS release, leading to endothelial membrane lipid peroxidation and smooth muscle cell (SMC) proliferation [7, 13]. Disrupted redox signaling alters vascular tone through redox-sensitive switches, contributing to vasoconstriction and endothelial dysfunction. Under OS, high peroxide levels inhibit COX-driven prostanoid synthesis, disrupting redox balance, promoting non-enzymatic peroxidation, and exacerbating oxidative damage following aSAH [3, 15, 63]. Combined with an inadequate endogenous antioxidant response, this creates a cycle of vascular and neural injury. Our findings, identifying LPMs that are associated with CVS and DCI, support the potential role of ROS in the onset of DCI.
Phosphatidylcholine, a key structural element of cellular membranes, is a substrate for generation of phosphatidylcholine hydroperoxide (PCOOH) when interacting with ROS [6, 39]. Cholesteryl ester hydroperoxide (CEOOH) forms under OS and serves as a marker of cholesterol ester oxidation and membrane lipid injury [25, 47, 68]. Suzuki et al. found significantly elevated PCOOH levels in cerebral arteries exhibiting vasospasm in a primate model of SAH [55]. Kamezaki et al. investigated the levels of PCOOH and CEOOH in CSF within 48 h of aSAH onset, showing that CEOOH levels were significantly associated with the subsequent development of symptomatic CVS (p = 0.002) [22]. While elevated PCOOH levels were not linked to CVS, they were independently associated with unfavorable 3-month functional outcomes, as measured by the Glasgow Outcome Scale (GOS), suggesting that PCOOH may reflect neuronal tissue damage rather than directly contributing to vasospasm [22]. Thus, PCOOH and CEOOH appear to have distinct, complementary roles in the early biochemical landscape following aSAH.
FFAs originate primarily from the hydrolysis of membrane phospholipids during cellular stress, and their accumulation contributes to oxidative damage. Acting as substrates for peroxidation and downstream mediators, FFAs initiate a cascade of detrimental effects following aSAH, including mitochondrial dysfunction and disruption of cerebrovascular tone [46]. For instance, oleic and linoleic acids reduce nitric oxide availability by diminishing calcium influx in endothelial cells, promoting cell dysfunction and endothelium-dependent constriction of SMCs [27, 28]. Highly unsaturated fatty acids like AA and DHA are particularly vulnerable to ROS-induced peroxidation, forming NPs and IsoPs that drive oxidative and inflammatory processes [40]. Pilitsis et al. demonstrated significantly elevated FFA levels in CSF within 48 h and again at 8 days post-aSAH, corresponding to early and delayed vasospasm phases. Patients who developed CVS showed early increases in linoleic, arachidonic, and palmitic acids, suggesting a link between lipid metabolism and vascular dysfunction [46]. While FFAs are not lipid peroxidation products themselves, their early rise reflects increased oxidative vulnerability and promotes the formation of secondary bioactive compounds, especially in hypoxic states caused by reduced cerebral blood flow.
COX- and LOX-derived metabolites play important roles in aSAH-related complications by promoting inflammation, endothelial dysfunction, and increased permeability. COX products, such as PGD2, PGF2α, PGE2, and TBXs, are primarily vasoconstrictive, while prostacyclin (PGI2) acts as a vasodilator [4]. LOX-derived leukotrienes (LTCs) further contribute to cerebral vessel constriction and brain edema [9, 44]. Maintaining a balance between these vasoconstrictors and vasodilators is crucial for preserving cerebral vascular tone, which supports the need for studying its disturbances in cerebrovascular pathologies. Oxidative stress and inflammation after aSAH disrupt this equilibrium, favoring vasoconstriction with elevated TBxA2 and diminished PGI2 levels [52]. Human studies have linked AA enzymatic oxidation metabolites to CVS, with Rodriguez y Baena et al. reporting higher CSF levels of PGD2, TBxB2, and LTC4 in symptomatic CVS patients in the early post-aSAH period [44, 49]. Similarly, Gaetani et al. noted increased levels of 6-keto-PGF₁α in patients with CVS, although the results did not reach statistical significance [14]. Our meta-analysis further suggests that early PGD₂ elevation may serve as a potential marker of vasospasm risk, with pooled data indicating that PGD₂ levels in CSF measured within three days post-aSAH were significantly higher in patients who developed CVS (MD = 537.24 pg/mL, p < 0.05) (Table 6). Additionally, elevated LTC levels have been associated with increased cerebral blood flow velocities measured via transcranial Doppler (TCD), reinforcing their link to vasospasm [59]. Ebselen, an LOX inhibitor and antioxidant, reduced vasospasm severity in SAH animal models and improved outcomes in ischemic stroke patients, which supports the role of LOX-derived products in exacerbating oxidative injury [17, 67]. A recent review by Solar et al. suggested that non-steroidal anti-inflammatory drugs (NSAIDs) targeting COX may outperform nimodipine in reducing CVS and provide complementary benefits when used in combination. However, clinical evidence remains limited as only one randomized controlled trial (RCT) reported improved outcomes and reduced vasospasm post-aSAH, while most supporting data comes from observational studies or preclinical models, limiting their clinical applicability [52]. Whether AA metabolites in CSF reflect a compensatory response to hemorrhage or act as agents driving vasospasm remains unclear. Gaetani et al. observed that AA metabolite concentrations increased over time in patients with CVS and correlated with the distribution of clotted blood in the subarachnoid space, indicating the importance of blood and hemoglobin release [14]. Clarifying such distinctions is crucial for understanding CVS pathophysiology. However, translating these findings to practice is challenging. Unlike systemic vascular diseases, changes in aSAH are confined within the blood-brain barrier (BBB) [2, 8, 66]. Although COX-derived metabolites can cross the BBB, NSAIDs commonly used in aSAH may inhibit their transport through the BBB, limiting their detectability in peripheral blood and reinforcing CSF as the preferred monitoring medium [57]. More human studies with standardized, time-resolved CSF sampling are needed to define the roles of these metabolites in CVS and DCI.
PLCs are formed through the ROS-induced oxidative modification of PUFAs. Elevated PLCs levels have been observed in CNS diseases associated with OS, including Alzheimer’s, Huntington’s, and Creutzfeldt-Jakob diseases [37, 41]. Among these, F2-IsoPs and F4-NPs, key products of non-enzymatic PUFA oxidation, are particularly reliable markers of oxidative damage. F2-IsoPs are derived from AA, found in all brain cells, while F4-NPs originate from DHA, a major component of neuronal phospholipids. Both are ROS-driven, COX-independent, and stable products, enhancing their reliability as biomarkers [40]. Studies have shown elevated F2-IsoPs in blood and CSF post-aSAH predicting DCI risk. Wiśniewski et al. reported that plasma F2-IsoPs measured on day 2 post-aSAH were associated with higher DCI risk, with an odds ratio (OR) of 1.04 (95% CI: 1.001–1.099; p = 0.04) [61]. Pan et al. also showed that plasma F2-IsoP levels above 0.51 ng/mL predicted poor 6-month functional outcomes post-aSAH with 82.8% sensitivity and 69.5% specificity [43]. However, studies suggest that variations in F2-IsoPs levels are more important in reflecting functional outcome when measured in CSF [33]. Wiśniewski et al. reported moderate predictive value for day 1 post-aSAH F2-IsoPs levels in CSF (AUC = 0.791, 95% CI: 0.619–0.963, p < 0.001) and stronger accuracy when combined with red cell distribution width (RDW) (AUC = 0.923, 95% CI: 0.825–1.000, p < 0.01) for identifying DCI risk [63]. Asaeda et al. reported that CSF levels of 8-iso-prostaglandin F₂α (8-iso-PGF₂α) - a representative F2-IsoPs - were significantly higher in patients with vasospasm on days 3–5 and 6–8 post-aSAH, with a peak observed between days 6 and 8 (p < 0.05), which aligns with the typical onset of CVS and suggests the concentration increase may be more related to subarachnoid clot presence than cerebral infarction [1]. Nevertheless, F2-IsoPs have shown significant associations with CVS and long-term outcomes when measured in urine (Table 6). Wiśniewski et al. in his studies reported that higher urinary F2-IsoPs concentrations on day 3 post-aSAH were associated with increased CVS risk, presenting with a sensitivity of 77.8% for detecting vasospasm [62]. Their findings also demonstrated elevated F2-IsoPs levels in patients with poor functional outcomes at 1, 4, and 12 months, as well as in those who developed acute hydrocephalus post-aSAH [60, 62]. These associations suggest that F2-IsoPs reflect the cumulative OS burden affecting vascular and neuronal integrity in aSAH and support their potential role in early prognosis and longitudinal monitoring. Therapeutically, lipid peroxidation inhibitors have also shown promise. Tirilazad was evaluated in a meta-analysis of five RCTs including 3,797 aSAH patients, where it reduced vasospasm incidence post-aSAH but did not improve functional outcomes, indicating a potential biological effect against vascular oxidative injury [20]. In contrast, edaravone, a free radical scavenger used in ischemic stroke and neurodegenerative diseases, improved functional outcomes in patients with brain ischemia by limiting ROS-induced injury [65]. Additionally, urinary measurement of LPMs provides a non-invasive, reproducible method for real-time tracking of OS, enabling risk stratification and treatment monitoring, while offering a practical alternative to invasive CSF or blood sampling.
F4-NPs may provide additional insights into neuronal OS and integrity, and some studies suggest that the increase in F4-NPs levels may hold greater predictive value than F2-IsoPs for early outcome prediction following aSAH [18]. Elevated F4-NPs in CSF shortly after bleeding were linked to post-aSAH complications, with Syta-Krzyżanowska et al. reporting F4-NPs as indicative of CVS on day 1 post-hemorrhage [56]. Hsieh et al. also found significantly higher F4-NP concentrations in CSF among patients who later exhibited poor clinical outcomes assessed by the GOS (p < 0.01), supporting a link between oxidative neuronal damage and long-term functional deficits [18]. Although few studies evaluate F4-NPs in aSAH complications, these findings suggest the importance of exploring their potential as biomarkers for aSAH-related complications.
Under conditions of high oxygen tension, oxidative injury shifts the balance of lipid peroxidation from F2-IsoPs to isofurans (IsoFs) formation [11, 16]. Therefore, the IsoFs to F2-IsoPs ratio may offer more insights into oxygen tension and mitochondrial dysfunction in aSAH. Gomes et al. showed elevated IsoFs to F2-IsoPs ratios in DCI patients, with IsoFs in CSF exhibiting 86% sensitivity for DCI detection [16]. However, the predictive value of IsoFs was based on samples taken on days 5–8 post-aSAH, which corresponds with the typical onset of CVS. This timing raises questions about whether IsoFs function as early predictors or as markers of an ongoing response to vasospasm and ischemia. Given the small sample size in the study, further prospective clinical validation is essential to better estimate the relevance of IsoFs for detecting DCI after aSAH.
Recent findings identify IAs as active vascular structures characterized by inflammation and tissue degeneration [5]. Starke et al. reviewed the role of ROS in IAs formation, highlighting their ability to induce endothelial injury, promote apoptosis, and recruit inflammatory cells [53]. Frösen et al. linked lipid accumulation, OS, and arterial wall degeneration, revealing that lipids and oxidized low-density lipoproteins (LDL) accumulate in the walls of saccular IAs, where they are ingested by mural cells and oxidized. This accumulation is strongly associated with mural cell loss and wall degeneration, often culminating in aneurysm rupture [13]. Hypertension-driven chronic inflammation further exacerbates redox imbalance, increasing ROS production and lipid oxidative damage, accelerating the weakening of the arterial wall, and potentially resulting in aneurysm formation [5]. Intriguingly, Frösen et al. observed that oxidized lipid levels in aneurysm walls did not correlate with systemic lipid profile of patients, suggesting a localized oxidative process independent of blood lipid disorders, which underscores the importance of studying oxidized lipid concentrations, even without the evidence for systemic dyslipidemia [13]. LPMs, as markers of OS, show promise as early indicators of arterial wall vulnerability, aiding in identifying patients at risk for IA development and rupture. Our review examines the potential association between LPMs and unruptured IAs, as reported by Syta-Krzyzanowska et al., who found that plasma F2-IsoPs and F4-NPs levels were three- and ten-fold higher in unruptured IA patients than controls [56]. These findings emphasize the need for further research into LPMs as biomarkers for IA screening, potentially paving the way for minimally invasive, blood-based techniques for early identification of individuals at high risk of IAs formation.
Traditional methods for monitoring DCI, such as clinical examinations and TCD, have significant limitations. TCD is operator-dependent, evaluates one cause of DCI, and offers moderate predictive value [26]. While clinical risk factors help stratify patients at risk for CVS and DCI, they do not guide prevention or treatment [64]. Diagnosing DCI in unconscious aSAH patients is particularly challenging, as up to 30% of lesions are clinically silent, which delays appropriate intervention [62]. Advanced monitoring techniques, such as cerebral perfusion flowmetry (CBF) or cerebral microdialysis (CMD), detect silent infarcts and metabolic disturbances but are invasive and not widely available [29, 62]. LPMs, such as F2-IsoPs, F4-NPs, and IsoFs, seem to hold potential for identifying patients at high risk of CVS and DCI. Validating their clinical utility requires large-scale prospective studies and standardized sampling protocols to assess their temporal dynamics. Diagnostic use also depends on establishing method-specific reference ranges and thresholds. Given inter-laboratory variability, mass spectrometry should be preferred for its sensitivity, specificity, and ability to quantify multiple metabolites simultaneously [32]. Reference values should be established based on healthy control populations to ensure consistency. Developing ISO-standardized calibrators and assays would further support the widespread adoption of LPMs as clinical tools. Simultaneous analysis of metabolites, such as F2-IsoPs and IsoFs, may provide further insights into OS intensity and oxygen tension through metabolite ratios [36]. Incorporating both vasoconstrictive and vasodilatory metabolites, such as TBXs, PGE2, and PGI2, into analysis can offer further understanding of the disrupted cerebrovascular balance in aSAH. Combining biomarker analysis with advanced imaging techniques may also provide a more comprehensive and dynamic understanding of CVS and DCI mechanisms. Efforts should address implementation of biomarker monitoring in routine clinical settings, exploring less invasive techniques, such as urine-based quantification, which presents a promising option for serial monitoring while reducing the burden on patients [60, 62]. Clinically, confirming the pathophysiological relevance of LPMs could pave the way for targeted therapies. Agents like tirilazad and edaravone already show biological activity in reducing oxidative injury in aSAH and ischemic stroke. Future research should investigate whether LPM-modulating agents can reduce the incidence or severity of vasospasm, enhance outcomes, or guide therapy effectiveness in aSAH by limiting oxidative damage. Future studies should aim to define cutoff values for specific LPMs and develop risk stratification nomograms for CVS and DCI, thereby maximizing their diagnostic and prognostic potential in routine aSAH management.
The predictive value of LPMs varies significantly across studies due to heterogeneity in patient selection, sample collection timing, and methodologies. Differences in inclusion and exclusion criteria contribute to variability in findings, while small effect sizes in several studies limit generalizability. Timing of sample collection represents a critical limitation, as it was inconsistently reported or standardized across studies, potentially obscuring the true predictive value of reported metabolites. Additionally, cutoff points, sensitivity, or specificity were not systematically reported in most studies, making it difficult to assess the reliability and consistency of findings. Varying DCI definitions add another layer of heterogeneity, affecting how metabolite levels are analyzed and compared across studies. Moreover, discrepancies in specimen collection, storage, and analysis may introduce additional variability, as lipid metabolites degrade rapidly if not handled properly [32]. Few studies date back to 1986–1988, raising concerns about the relevance of their findings given advancements in analytical techniques and clinical management of aSAH, which may limit their clinical applicability today. An additional source of methodological variability lies in the analytical techniques used to quantify LPMs. While some studies relied on enzyme-linked immunosorbent assays (ELISA), others used methods such as liquid chromatography–mass spectrometry (LC-MS) or gas chromatography–mass spectrometry (GC-MS), which differ significantly in terms of sensitivity, specificity, and reproducibility. These issues, along with failure to control for other factors such as coexisting conditions in many studies, highlight the need for standardized sampling protocols and better matching of patients and controls in future studies. Addressing these challenges requires large-scale, prospective studies with harmonized protocols for patient selection, sample handling, standardized DCI definitions, and clinical endpoints to establish LPMs as reliable biomarkers in aSAH.
This review highlights the potential of enzymatic and non-enzymatic lipid metabolites as biomarkers for OS and complications like CVS and DCI in aSAH. F2-IsoPs and AA derivatives, such as 6-keto-PGD-F1-α, PGD2, and LTC4, show promise for predicting early DCI risk, particularly within the first three days post-aSAH, aiding clinical prognosis and timely intervention. Variability in patient selection, sample timing, and analytical methods emphasizes the need for standardized protocols. Non-invasive options like urine-based F2-IsoPs monitoring present promising alternatives to CSF sampling. Large-scale prospective studies are essential to confirm LPMs’ predictive value and validate their clinical thresholds. This review highlights the significance of lipid disturbances in aSAH and calls for harmonized research designs to advance CVS and DCI biomarker development.
Below is the link to the electronic supplementary material.
Supplementary Material 1