Introduction

Aneurysmal subarachnoid hemorrhage (aSAH) is a life-threatening condition with a high mortality rate of 50%, and even patients receiving optimal medical care may experience long-term disability1,2,3. As the global number of individuals who suffer new strokes, die, or remain disabled after a stroke increases, early identification of prognostic risk factors is crucial for targeted prevention4.

Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) are commonly used to assess hepatic function in clinical practice. ALT is more liver-specific, while AST levels can rise due to ischemic cell death in various nonhepatic tissues such as the kidneys, heart, and brain. The AST to ALT ratio was first introduced by De Ritis et al. in 1957 to differentiate viral hepatitis from other icteric and anicteric hepatic diseases5. Accumulating evidence suggests that an increased AST to ALT ratio (AAR, De Ritis) may serve as a surrogate marker for ischemic end-organ damage, correlating with higher rates of cardiovascular events and mortality6. The underlying pathological effects of elevated De Ritis on cardiovascular organs include mitochondrial dysfunction and oxidative7,8.

Since the De Ritis ratio has demonstrated prognostic value in various cancer entities, some research groups have explored its clinical significance in neurosurgical settings. Christoph Wetz et al. demonstrated that a high De Ritis ratio improved the prediction of progression-free survival after neuroendocrine tumor treatment but not during the treatment itself9,10,11. Pre-radiotherapy and postoperative De Ritis levels were revealed as novel serum prognosticators in newly diagnosed atypical meningiomas and glioblastomas, respectively12,13. In the neurovascular department, the De Ritis ratio may serve as a guide for intraoperative transfusion in unruptured intracranial aneurysm surgery14. Furthermore, a dynamic fluctuation in the De Ritis ratio provided a significant independent prediction for mortality in patients with moderate-to-severe traumatic brain injuries15.

To our knowledge, no prior studies have considered using De Ritis to predict the prognosis of ruptured aneurysms or incorporated this ratio into an established risk model. However, the relationship between De Ritis and mortality, as well as long-term functional outcomes after aSAH, has not yet been sufficiently elucidated. This study aimed to explore the association between De Ritis and clinical outcomes after aSAH, including poor functional outcomes at discharge and 90 days, as well as in-hospital complications.

Methods

Study design

We retrospectively reviewed the patient data from consecutive aSAH patients admitted to our institution between January 2015 and September 2022. All patient data were from the Long-term Prognosis of Emergency Aneurysmal Subarachnoid Hemorrhage (LongTEAM, ClinicalTrials.gov Identifier: NCT04785976) registry. This study was approved by the institutional review board of Beijing Tiantan Hospital. Informed consent for clinical analyses was obtained from all individual participants or their authorized representatives, and all the analyses were performed in accordance with the Declaration of Helsinki and the local ethics policies. Both procedures were performed by specific senior neurosurgeons, with an annual average of more than 300 procedures per neurosurgeon. All patients were managed according to the guidelines from the American Heart Association/American Stroke Association and institutional routine.

Inclusion and exclusion criteria

All patients had angiographically documented aSAH confirmed by CT or lumbar puncture. In this study, the inclusion criteria were 1) age ≥ 18 years; 2) emergency admission; 3) no previous aneurysm rupture; 4) only patients treated by surgical clipping or endovascular treatment; 5) less than 72 h from rupture to admission and less than 72 h from admission to treatment; 6) single aneurysm; and 7) complete 90-day follow-up. Exclusion criteria were 1) other neurological diseases (tumor, vascular malformation, Parkinson’s disease, multiple sclerosis, and primary epilepsy) and functional or neurological deficit of the extremities due to any cause; 2) history of neurosurgery prior to rupture; and 3) treatment, including external ventricular drainage, intubation, and/or mechanical ventilation, at other hospitals before presentation to our hospital.

Data collection

Baseline clinical characteristics and imaging data were reviewed, such as age, sex, body mass index (BMI; calculated as weight in kilograms divided by the square of the height in meters, kg/m2), location of the ruptured aneurysm, intraventricular hemorrhage, acute hydrocephalus, and medical history (e.g. hypertension, diabetes mellitus, hyperlipidemia, stroke and heart disease). The World Federation of Neurosurgical Societies (WFNS) grade, modified Fisher Scale (mFS) grade, Graeb score, Subarachnoid Hemorrhage Early Brain Edema Score (SEBES), and modified Rankin Scale (mRS) score were assessed. Postoperative clinical complications during hospitalization were collected, including rebleeding, delayed cerebral ischemia (DCI), intracranial infection, stress ulcer bleeding, abnormal hepatic function, urinary tract infection (UTI), anemia, hypoproteinemia, pneumonia, and deep vein thrombosis (DVT), electrolyte disturbance, and disorders of lipoprotein metabolism. Supplemental Table 1 shows the detailed diagnostic criteria for in-hospital complications. The mRS score and mortality rate were collected at discharge.

Table 1 Baseline characteristics according to quartiles of De Ritis.
Full size table

Exposure

To mitigate surgical and endovascular interference, we designated postoperative De Ritis as the exposure variable. Fasting blood samples were collected within 24 h after treatment routinely. Both preoperative and postoperative serum AST and ALT were tested immediately using standard methods at the laboratory. Laboratory tests including WBC (white blood cell), total protein (TP), albumin (ALB), globulin (GLB), total bilirubine (TBIL), total cholesterol (TC), triglyceride (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), apoprotein A1 (APO_A1), and apoprotein B (APO_B) were measured and documented. For normal tests values, please see the Supplementary Table 2.

De Ritis was calculated as AST divided by ALT. Participants were classified into four groups according to quartiles of De Ritis and two groups by the cut-off value of 1 according to previous studies16,17.

Outcome assessment

The primary outcome was the mRS (a stroke outcome scale with scores ranging from 0 [no symptoms] to 6 [dead]) score at discharge and 90 days after discharge. The certified clinical research coordinators and neurosurgeons followed up with patients via telephone or an outpatient appointment 90 days after discharge. Unfavorable outcome was defined as an mRS score of 3 to 6 or 4 to 6. Second outcomes were defined as presence of prespecified in-hospital complications.

Statistical analysis

The descriptive statistics are summarized as median (interquartile range, IQR) for continuous variables with skewed distribution and frequency (percentage) for categorical variables. After testing for normality, continuous variables were analyzed using the nonparametric Mann–Whitney U test or Kruskal–Wallis test. The Pearson chi-square test, continuity correction test, or Fisher’s exact test were used to test the dichotomized and categorical independent variables.

For poor functional outcome and in-hospital complications, the logistic regression model and OR with 95% CI was used18. Patients with the lowest quartile of De Ritis or De Ritis < 1 were defined as the reference group. Three adjusted models were fitted. Model 1 adjusted for age (continuous) and sex. Model 2 further adjusted for BMI, current smoker, current alcohol drinking, medical history (hypertension, diabetes mellitus, dyslipidemia, stroke, heart disease), the WFNS, mFS score at admission, location of aneurysm and treatment modality. In model 3, laboratory tests including WBC (continuous), TC, TG, HDL, and LDL were further adjusted17. Trend tests were performed using quartiles as ordinal variables. To assess the incremental predictive value of De Ritis in addition to conventional risk factors, Delong test, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were applied. In addition, restricted cubic spline models were used to assess the dose–response relationships between De Ritis and clinical outcomes, and the knots were set at the lowest akaike information criterion (AIC) value for a better quality of model fitting.

To examine the robustness of our findings, we further did sensitivity analysis by excluding patients with De Ritis > 2, who may have underlying liver disease.

All statistical analyses were performed using R version 4.4.0 Statistical Software, GraphPad PRISM 8.3.0 (GraphPad Software Inc.) and validated by SPSS Statistics version 29.0 (IBM Corp.). Statistical significance was set at a two-tailed P < 0.05.

Results

Patient characteristics

After exclusion of 170 ineligible patients, a total of 1,098 patients in the LongTEAM cohort who had their hospitalization between January 2015 and September 2022 were enrolled in the present study. The baseline characteristics between the included patients and those excluded are presented in Supplemental Table 3. Patients excluded from the current study were more less likely to be smoker, alcohol drinker, had lower proportion of diseases history (such as hypertension, hyperlipidemia, history of stroke, and onset-loss of consciousness).

Among the 1,098 patients, the median age was 55 years (IQR, 48 to 63), 58.5% were female, 24.6% were current smoker, and 18.4% were current alcohol drinkers. The median De Ritis was 1.20 (IQR, 0.89 to 1.58). The demographics and clinical characteristics of the study population are shown in Table 1. As compared to patients with the lowest De Ritis quartile, those with higher De Ritis level were older, more likely to be women, with loss of consciousness, WFNS 4–5, and underwent surgical clipping treatment, less likely to be obesity, current smokers or drinkers. Moreover, patients with higher De Ritis level had high postoperative HDL and AST concentration, but lower preoperative ALT, AST, postoperative total protein, albumin, TG, LDL, AST, Apoprotein A1 and Apoprotein B concentration.

De Ritis and functional outcomes

At discharge assessment, there were 446 (40.6%) patients with mRS 3–6, 234 (21.3%) patients with mRS 4–6, and 17 (1.5%) had died. During the 3 months assessment, there were 210 (19.1%) patients with mRS 3–6, 131 (11.9%) patients with mRS 4–6, and 28 (2.6%) had died (Table 2).

Table 2 Associations of De Ritis with poor functional outcomes at discharge and 3 months follow-up.
Full size table

After adjusting for all the potential covariates (Model 3), patients in the second De Ritis quartile had a lower risk of unfavorable functional outcome (both mRS 3–6 and mRS 4–6) at discharge assessment (OR 0.623, 95% CI 0.408–0.951; OR 0.547, 95% CI 0.318–0.943 respectively), compared with those in the reference quartile. We also observed increased risks of mRS 3–6 and mRS 4–6 at 90 days for patients in the fourth De Ritis quartile in age and sex adjusted model, but the risks were attenuated and became insignificant after adjusted all covariates. We also observed a significant linear association between De Ritis and death at both discharge and 90 days (P for trend = 0.018, P for trend = 0.016, respectively) after adjusted for all covariates.

Restricted cubic spline models showed a non-linear associations between De Ritis levels and risk of mRS 4–6 at discharge assessment (Table 3, Fig. 1A). However, for mRS 3–6 and death at discharge and 90 days assessment, the RCS model failed to reach statistical significance (Fig. 1B).

Table 3 Associations of De Ritis with poor functional outcomes at discharge and 3 months follow-up.
Full size table
Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Non-linear associations between De Ritis ratio and clinical outcomes in patients with aneurysmal subarachnoid hemorrhage (aSAH). Restricted cubic spline regression models with fully adjusted covariates were used to explore the dose–response relationship between postoperative De Ritis levels and the risks of (A) unfavorable outcome (modified Rankin Scale score 4–6) at discharge, (B) death at discharge, (C) major adverse cardiac events (MACE), (D) abnormal hepatic function, (E) anemia, and (F) disorders of lipoprotein metabolism. Solid lines represent adjusted odds ratios (ORs) and shaded areas indicate 95% confidence intervals. Models were adjusted for age, sex, BMI, smoking, alcohol drinking, medical history, WFNS grade, modified Fisher score, aneurysm location, treatment modality, and laboratory variables including WBC, TC, TG, HDL, and LDL.

Incremental effect of De Ritis in predicting 90 days unfavorable outcomes

In an analysis of 90 days unfavorable outcomes (mRS 3–6), ROC curves were constructed to assess the predictive ability of the TAPS prognostic model19 and the TAPS model plus De Ritis index respectively. There was no significant difference between the TAPS model (AUC: 0.801) and the De Ritis index model (AUC: 0.801) (P = 0.345). No improvement was observed in the predictive power of the TAPS model by incorporating the De Ritis index in patients with aSAH (net reclassification improvement [NRI]: 0.0022 [−0.0409–0.0384], P = 0.9457; integrated discrimination improvement [IDI]: 0.0183 [−0.0008–0.0007], P = 0.8522).

De Ritis and in-hospital complications

The three most prevalent in-hospital complications were electrolyte disturbance (813/1098 [74.0%]), hypoproteinemia (403/1098 [36.7%]), and MACE (392/1098 [35.7%]).

After adjusting for all the potential covariates (Model 3), patients in the second De Ritis quartile had a higher risk of MACE (OR 1.550, 95% CI 1.039–2.312), and third quartile a higher risk of DCI (OR 1.569, 95%CI 1.047–2.353), and fourth quartile with a higher risk of anemia (OR 1.662, 95%CI 1.092–2.530), compared with those in the reference quartile. Furthermore, patients with higher levels of De Ritis had lower risk of hepatic function (Q2, Q3, and Q4) and disorder of lipoprotein metabolism (Q3 and Q4) in fully-adjusted models. We found that the risk of anemia was significantly increased with increasing De Ritis (P for trend = 0.017).

Restricted cubic spline models showed a non-linear associations between De Ritis levels and risk of MACE (knot = 5), abnormal hepatic function (knot = 6), anemia (knot = 5) and disorders of lipoprotein metabolism (knot = 3, Table 3, Fig. 1C-1F). However, no other in-hospital complications showed statistical significance.

Sensitivity analysis

When patients were classified into two groups, those with De Ritis ≥ 1 had a higher risk of anemia (aOR 1.457, 95%CI 1.070–2.031, Table 4) and a lower risk of abnormal hepatic function (aOR 0.596, 95%CI 0.434–0.818) and disorders of lipoprotein metabolism (aOR 0.722, 95%CI 0.523–0.997), consistent with the findings with quartiles. Similar findings were robust after excluding participants with De Ritis > 2 in sensitivity analysis (Table S4 and Table S5).

Table 4 Associations of De Ritis with in-hospital complications.
Full size table

Discussion

The primary finding of the current study conducted within the LongTEAM registry was that elevated De Ritis levels were associated with an increased risk of unfavorable functional outcomes at discharge and several in-hospital complications in patients with aSAH. These associations persisted even after full adjustment for potential confounders. However, there was no supplementary incremental performance of De Ritis for the aSAH prognostic TAPS model.

Large aSAH cohort studies have demonstrated worse clinical outcomes in patients with chronic liver disease compared to those without, and a higher Model for End-Stage Liver Disease score in aSAH patients was linked to an increased risk of mortality20,21. De Ritis serves as a readily available non-invasive marker of hepatocellular damage and liver disease22. Recent studies have indicated that an association of De Ritis ratio with an increased risk of cardiovascular diseases, such as acute type A aortic dissection23, acute heart failure24, acute ischemic stroke25, and other cardiovascular diseases. Furthermore, several large clinical studies have reported that an elevated De Ritis ratio was independently associated with an unfavorable prognosis and mortality, even after adjusting for other well-established vascular risk factors26,27,28. Consistent with previous findings that increased De Ritis is associated with worse functional outcomes, our results indicated a positive correlation between De Ritis and higher mortality rates in aSAH patients (P for trend = 0.016).

There is substantial evidence that liver enzymes play a crucial biological role in coronary artery calcification, endothelial dysfunction, and metabolic disorders, all of which are critical determinants of clinical outcomes in stroke29. Chun-On Lee conducted a National Health and Nutrition Examination Survey from 1999 to 2016 and found that while liver fat score was not associated with myocardial infarction or stroke, it was linked to increased all-cause and cardiovascular mortality, with hazard ratios (HRs) of 1.10 (95% CI: 1.07–1.13) (p < 0.001) and 1.12 (95% CI: 1.06–1.17) (p < 0.001), respectively29. However, to the best of our knowledge, no study has assessed the association between De Ritis and the prognosis of hemorrhagic stroke. In our study of 1,098 patients, a higher postoperative De Ritis ratio was associated with poor functional outcomes (defined as mRS 3–6) at discharge. We also observed a non-marginal significance at 3 months in the fully adjusted model (P = 0.398). The discrepancy with previous results from ischemic stroke may be due to short-term strong prognostic factors overshadowing the role of De Ritis. Several plausible explanations exist for the association between De Ritis and poor outcomes after aSAH. First, postoperative complications such as MACE, DCI, and anemia are significant, and De Ritis can predict these events to some extent. Second, previous studies have found that excessive glutamate is released from neurons into the brain parenchyma extracellular space after a stroke. A small increase in glutamate concentration in the cerebrospinal fluid can cause a pronounced increase in intracellular calcium and trigger neuronal death. Therefore, glutamate is associated with greater stroke severity, larger infarct volume, and worse functional outcomes. Although the direct roles of ALT and AST in the regulation of glutamate are less known, the increase in CSF glutamine may result from increased activity of aspartate aminotransferase, an enzyme that forms glutamine, in the pathological brain30. High glutamate levels within the lesion, monitored by cerebral microdialysis, are linked to a greater likelihood of DCI occurrence31. In our study, we discovered that the third quartile of the De Ritis ratio independently correlated with DCI, even after adjusting for other significant predictors, indirectly aiding in the early management of 3-month mRS in patients. Third, our results indicated a statistical relevance between the second quartile of De Ritis and postoperative MACE, as well as a nonlinear relationship. These findings are clinically significant because aSAH patients at increased risk of MACE could be identified using the easily accessible and limited threshold of the aminotransferase ratio.

Given that the liver receives nearly a quarter of the total cardiac output, the relevance of De Ritis to systemic blood flow changes is logical and it responds swiftly. Recent studies have demonstrated that De Ritis has a strong prognostic role in predicting both CVD incidence and cardiovascular mortality over more than a decade of follow-up in large cohorts from European and Asian populations32,33. De Ritis is noted as a potent predictive and indicative biomarker of metabolic syndrome34. Our study suggests that the smaller the AST to ALT ratio, the more pronounced the disorders of lipoprotein metabolism. In cases positive for metabolic syndrome, the De Ritis ratio tends to be lower, consistent with previous studies35. Further research is needed to investigate the potential of these biomarkers and their appropriate reference values to more accurately predict aSAH population risks, which may be particularly valuable for risk stratification.

The primary strength of our study lies in its extensive aSAH registry, comprising over 1,000 participants, providing ample statistical power. To mitigate the influence of outliers, we excluded patients with a De Ritis ratio greater than 2 and discovered similar conclusions. Additionally, for practical clinical application, we calculated the study results with a critical value of 1, arriving at the same conclusion.

Limitations

Our study should be considered in light of certain limitations. First, the De Ritis index was not initially developed for cerebrovascular disorder populations. We emphasize the prognostic role of asymptomatic elevation of De Ritis among stroke patients with suspected lipoprotein disorder risks. Second, the specific threshold of De Ritis contributing to adverse outcomes after aSAH requires further validation. Although many covariates were adjusted in this study, the possibility of residual confounding cannot be ruled out. Additionally, since repeated measurements of De Ritis were unavailable, we cannot conclude the cumulative and causal relationship between De Ritis and clinical outcomes of aSAH. Third, this study cohort was derived from a single center in China, potentially limiting the applicability of the findings to other ethnic groups or populations with different medical systems. Fourth, only a 90-day follow-up was considered in the analysis; longer-term outcomes related to De Ritis should be investigated in the future.

Conclusions

In patients with aSAH, elevated De Ritis levels were correlated with an increased risk of unfavorable functional outcomes at discharge, MACE, DCI, and anemia. Recognizing that postoperative De Ritis levels may aid in identifying high-risk aSAH patients, these findings hold significant clinical relevance.