Abstract
Objectives:
The aim of this study was to evaluate the diagnostic accuracy of the cerebrospinal fluid glucose-to-lactate ratio (CGLR) for detecting postoperative intracranial infections in patients with acute brain injury (ABI) and compare it with conventional biomarkers.
Methods:
A single-center prospective study was conducted with 121 postoperative ABI patients suspected of infection. Cerebrospinal fluid (CSF) and peripheral blood samples were collected under aseptic conditions. Levels of CSF lactate (cLac), CSF glucose (cGlu), CSF procalcitonin (PCT), and CGLR were measured.
Results:
The AUC values for cGlu and cLac were 0.836 and 0.820, respectively, with optimal cutoffs of 2.6 mmol⋅L–1 for cGlu and 4.5 mmol⋅L–1 for cLac. The combined metric, CGLR, demonstrated a higher diagnostic performance (AUC = 0.866) with an optimal cutoff of 0.61. CGLR values were similar between hyperglycemic and normoglycemic patients, regardless of infection status (all P > 0.05). The diagnostic performance of CGLR was consistent across SAH, ICH, and TBI patient cohorts (all P > 0.05).
Conclusion:
CGLR is a rapid, cost-effective, and reliable biomarker for diagnosing postoperative intracranial infections in ABI patients. It outperforms traditional markers and remains unaffected by peripheral glucose fluctuations.
1 Introduction
Acute brain injury (ABI), encompassing subarachnoid hemorrhage (SAH), intracranial hemorrhage (ICH), and traumatic brain injury (TBI), is characterized by high mortality and disability rates. ABI represents a frequently encountered severe condition in neurosurgery resulting from head trauma (). Surgical intervention is often lifesaving (). However, postoperative intracranial infection remains a major challenge in ABI management and a critical risk factor for mortality (–). Early detection of intracranial infection is essential. Although cerebrospinal fluid (CSF) bacterial culture is the standard criterion for diagnosing postoperative infection, it has limitations, including susceptibility to contamination, prolonged processing times, and low positivity rates (). Thus, identifying CSF biomarkers that reliably predict postoperative intracranial infection could enable timely treatment.
CSF lactate (cLac), CSF glucose (cGlu), and CSF procalcitonin (cPCT) have demonstrated usefulness in diagnosing postoperative intracranial infection (–). Determining cLac levels is cost-effective, easily accessible, and unaffected by leukocyte or erythrocyte presence in CSF (). Although the CSF glucose-to-lactate ratio (CGLR) has recently been associated with ABI prognosis, its relationship with postoperative intracranial infection remains unclear (, ).
This study aims to evaluate the reliability of CGLR as a biomarker for postoperative intracranial infection following ABI and compare its diagnostic performance with traditional markers (cLac, cGlu, and CSF PCT).
2 Materials and methods
2.1 Study design and participants
This prospective observational study was conducted at a single center (Intensive Care Unit, People’s Hospital of Ningbo University) from June 2024 to June 2025. Patients with ABI who had undergone surgery and subsequently exhibited fever and suspicion of intracranial infection were included. CSF samples (obtained via lumbar puncture) and serum samples were collected within 24 h. Exclusion criteria included: (1) diagnosis of sepsis, (2) history of other neurosurgical procedures, (3) contraindications to lumbar puncture, (4) renal failure, (5) age under 18 years, (6) pregnancy, (7) hypoxic-ischemic encephalopathy, and (8) epilepsy. Each sample was collected from a unique patient, with no repeated sampling from the same individual. The study complied with the Declaration of Helsinki (2000) and was approved by the Ethics Committee of the Affiliated People’s Hospital of Ningbo University (2024058).
2.2 Diagnostic criteria
Diagnostic criteria for intracranial infection were based on the CDC/NHSN Surveillance Definitions (). Patients were diagnosed with intracranial infection if they met either of the following conditions: (1) identification of a pathogen through positive CSF culture or confirmation via next-generation sequencing (NGS); or (2) presence of at least one clinical symptom not attributable to other causes (e.g., temperature above 38°C, meningeal irritation signs, focal neurological deficits, or headache) along with at least one of the following: positive blood culture, positive CSF Gram staining, or abnormal CSF parameters (white blood cell count > 100 cells/mm3, protein > 50 mg/dL, and glucose < 2.5 mmol/L).
In this study, CSF underwent next-generation sequencing (NGS) with approval from patients or their families.
2.3 Sample collection
Causes of ABI, including TBI, ICH, and SAH, were recorded. Basic demographic data (age, gender) and clinical findings were also collected. Levels of leukocytes, lactate, glucose, and PCT were measured in serum and CSF for both groups.
CSF samples were collected from all participants by lumbar puncture. The GEM Premier 4,000 system (United States) was used as an arterial blood gas analyzer to measure lactate and glucose levels. The enzyme-linked fluorescence method (Marcy L’Etoile, France) was used to determine PCT concentrations, with a minimum detection threshold of 0.05 ng/mL. Under sterile conditions, peripheral venous blood samples were collected within 30 min after obtaining CSF.
All CSF samples were analyzed within 30 min of collection and stored at 4°C until analysis. No freeze-thaw cycles occurred prior to glucose and lactate measurements.
2.4 Statistical analysis
Based on a preliminary CGLR AUC of 0.85 from a pilot study of 20 patients, a minimum sample size of 90 patients was required (power of 0.8, alpha of 0.05). To account for potential exclusions, 121 patients were enrolled.
Statistical analysis was performed using SPSS software (version 24.0, IBM Corporation, Armonk, NY, United States). Continuous variables with normal distribution are expressed as mean ± standard deviation (SD), while skewed data are presented as median (interquartile range, IQR). Categorical variables are presented as numbers (percentages). Differences between groups were assessed using the Chi-square test for categorical variables. The Mann-Whitney U test or Student’s t-test were used for continuous variables. ROC curves were constructed to evaluate diagnostic performance. P < 0.05 were considered statistically significant. The statistical analysis plan was finalized prior to data collection and remained unchanged after data access.
3 Results
3.1 Study population
A total of 121 CSF samples were collected during the study period. Of these initial samples, 7 were excluded due to contamination or equivocal culture results. Another 22 samples were classified as diagnostically inconclusive. The remaining 92 samples, clearly defined as either intracranial infection (n = 41) or no infection (n = 51), were included in the final analysis (Figure 1). Clinical characteristics of enrolled participants are presented in Table 1.
FIGURE 1
TABLE 1
| Characteristics | Intracranial infection (n = 41) | No Intracranial infection (n = 51) | P-value |
|---|---|---|---|
| Ages (year) | 59.5 ± 13.30 | 57.8 ± 15.56 | 0.584 |
| Sex (male) | 26 (63.41%) | 34 (66.67%) | 0.745 |
| Type of diseases | |||
| TBI | 13 (31.71%) | 17 (33.33%) | 0.585 |
| ICH | 21 (51.22%) | 26 (50.98%) | 0.560 |
| SAH | 7 (17.07%) | 8 (15.69%) | 1.000 |
| Clinical findings | |||
| Diabetes mellitus | 10 (24.39%) | 21 (41.18%) | 0.088 |
| On dexamethasone | 4 (9.76%) | 7 (13.73%) | 0.565 |
| GCS | 6.10 ± 1.59 | 6.61 ± 1.56 | 0.126 |
| Blood parameters | |||
| WCC (×109/L) | 10.36 ± 4.80 | 11.35 ± 3.90 | 0.367 |
| PCT (ng/ml) * | 0.15 (0.10–0.29) | 0.15 (0.08–0.42) | 0.778* |
| bGlu (mmol/L) * | 8.40 (7.40–11.10) | 9.90 (6.80–13.7) | 0.646* |
| bLac (mmol/L) * | 1.40 (1.00–1.85) | 1.50 (1.10–2.10) | 0.431* |
| CSF parameters | |||
| cLac (mmol/L) * | 5.80 (4.10–7.75) | 3.20 (2.20–4.30) | < 0.001* |
| cGlu (mmol/L) * | 2.10 (1.65–3.30) | 4.40 (3.30–5.50) | < 0.001* |
| cPro (mg/L) * | 2504.00 (1196.50–3387.50) | 946.75 (535.50–2220.25) | 0.001* |
| CSF leukocyte (/uL) * | 875.00 (150.00–2868.00) | 55.00 (5.00–467.50) | < 0.001* |
| CSF PCT (ng/ml) * | 0.17 (0.12–0.28) | 0.13 (0.095–0.21) | 0.030* |
| CGLR* | 0.38 (0.24–0.60) | 1.45 (0.81–2.12) | < 0.001* |
Characteristics of samples stratified according to confirmed intracranial infection and no intracranial infection groups.
TBI, traumatic brain injury; ICH, intracranial hemorrhage; SAH, subarachnoid hemorrhage; GCS, Glasgow Coma Scale; WCC, white cell count; PCT, procalcitonin; bGlu, blood glucose level; bLac, blood lactate level; CSF, cerebrospinal fluid; cLac, cerebrospinal fluid lactate; cGlu, cerebrospinal fluid glucose; cPro, cerebrospinal fluid protein; CGLR, cerebrospinal fluid glucose-to-lactate ratio; IQR, interquartile range; SD, standard deviation. *Data presented in median (IQR). The rest of data presented in n (%) or mean SD.
Microbiological findings are summarized in Table 2. Among the 41 infected patients, 17 (41.5%) had positive CSF culture or NGS results. Monomicrobial infections were identified in 10 cases (58.8%), and polymicrobial infections in 7 cases (41.2%).
TABLE 2
| Category | n | Proportion (%) |
|---|---|---|
| Total infected patients | 41 | 100 |
| Culture/NGS positive | 17 | 41.5 |
| Culture/NGS negative | 24 | 58.5 |
| Among culture/NGS positive (n = 17) | ||
| Monomicrobial infections | 10 | 58.8 |
| Polymicrobial infections | 7 | 41.2 |
| Monomicrobial infections (n = 10) | ||
| Acinetobacter baumannii | 4 | 40.0 |
| Stenotrophomonas maltophilia | 2 | 20.0 |
| Roseomonas mucosa | 1 | 10.0 |
| Staphylococcus haemolyticus | 1 | 10.0 |
| Staphylococcus caprae | 1 | 10.0 |
| Chlamydia pneumonia | 1 | 10.0 |
| Polymicrobial infections (n = 7) | ||
| Enterococcus faecium + Staphylococcus hominis | 1 | 14.3 |
| Acinetobacter baumannii + Aspergillus fumigatus | 1 | 14.3 |
| Porphyromonas gingivalis + Parvimonas micra + Fusobacterium nucleatum | 1 | 14.3 |
| Pseudomonas aeruginosa + Acinetobacter baumannii | 1 | 14.3 |
| Klebsiella pneumoniae + Cytomegalovirus | 1 | 14.3 |
| Streptococcus pneumoniae + Klebsiella pneumoniae | 1 | 14.3 |
| Corynebacterium accolens + Candida parapsilosis | 1 | 14.3 |
Microbiological findings in patients with postoperative intracranial infection (n = 41).
No significant differences existed between infected and non-infected groups regarding age, gender, disease type, clinical findings, or blood parameters (P > 0.05). However, significant differences emerged in CSF parameters. Levels of cPro, cLac, and CSF leukocytes were significantly elevated in the intracranial infection group (P < 0.005). Additionally, CSF PCT concentrations were higher in infected patients (P < 0.05). The intracranial infection group also showed significantly reduced CGLR and cGlu (P < 0.001) (Figure 2 and Table 1).
FIGURE 2
3.2 Comparison of systemic (bGlu, bLac) and central (cGlu, cLac, CGLR) biomarkers
A weak but significant correlation existed between glucose levels in CSF and blood (r = 0.267, P = 0.01), while no significant association was observed for lactate (r = 0.071, P = 0.502) (Figures 3A–D). Meanwhile, cLac levels showed a moderate negative correlation with cGlu (Spearman r = -0.560, P < 0.001), whereas a minimal association appeared between CSF PCT and cLac (Figures 3B,C).
FIGURE 3
Subgroup analysis revealed a weak negative correlation without statistical significance between lactate and CSF glucose in patients without infection (r = -0.259, P = 0.069) (Figure 3E). In contrast, patients with infection showed a strong, significant negative correlation for these parameters (r = -0.549, P < 0.001) (Figure 3F).
Patients were stratified into normoglycemic and hyperglycemic subgroups using a blood glucose cutoff of 11.1 mmol/L (). No significant differences appeared in CGLR between normoglycemic and hyperglycemic subgroups, regardless of intracranial infection (both P > 0.05). However, CGLR was significantly lower in infected patients compared to non-infected patients within each subgroup (both P < 0.001) (Figure 4).
FIGURE 4
ROC analysis showed an AUC of 0.836 for cGlu with an optimal cutoff of 2.6 mmol/L (Figure 5A and Table 3). cLac had an AUC of 0.820 at a cutoff of 4.5 mmol/L. The CSF-to-blood glucose ratio (CBGR) showed moderate diagnostic accuracy (AUC 0.781, Youden index 0.45), whereas the combined lactate-to-glucose ratio (CLGR) demonstrated good accuracy (AUC 0.832, Youden index 0.58). CGLR had the best diagnostic performance, with the highest AUC (0.866), specificity (90.0%), and Youden index (0.68) among all biomarkers.
FIGURE 5
TABLE 3
| CSF Maikers | AUC(95% CI) | Cut-off point | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Youden index |
|---|---|---|---|---|---|---|---|
| cLac (mmol/L) | 0.820 (0.726–0.893) | >4.5 | 73.17 | 82.00 | 76.9 | 78.8 | 0.55 |
| cGlu (mmol/L) | 0.836 (0.744–0.905) | ≤ 2.6 | 63.41 | 96.00 | 92.9 | 76.2 | 0.59 |
| cPro (mg/L) | 0.710 (0.600–0.804) | >1,079 | 80.56 | 55.32 | 58.0 | 78.8 | 0.36 |
| CSF leukocyte (/uL) | 0.756 (0.652–0.843) | >110 | 78.38 | 63.27 | 61.7 | 79.5 | 0.42 |
| CSF PCT (ng/ml) | 0.631 (0.521–0.732) | >0.22 | 41.03 | 83.33 | 66.7 | 63.5 | 0.24 |
| CBGR | 0.781 (0.683–0.861) | ≤ 0.29 | 60.98 | 84.00 | 75.8 | 72.4 | 0.45 |
| CLGR | 0.832 (0.740–0.903) | >12.32 | 80.49 | 78.00 | 75.0 | 83.0 | 0.58 |
| CGLR | 0.866 (0.779–0.929) | ≤ 0.61 | 78.05 | 90.00 | 86.5 | 83.3 | 0.68 |
ROC curve analysis results of CSF marker in differentiating intracranial infection and no intracranial infection.
cLac, cerebrospinal fluid lactate; cGlu, cerebrospinal fluid glucose; cPro, cerebrospinal fluid protein; CSF, cerebrospinal fluid; PCT, procalcitonin; CBGR, cerebrospinal fluid glucose-to-blood glucose ratio; CLGR, cerebrospinal fluid lactate-to-CBGR ratio; CGLR, cerebrospinal fluid glucose-to-lactate ratio.
Additional subgroup analyses were performed. CGLR exhibited superior diagnostic accuracy in patients with SAH, ICH, and TBI, with corresponding AUC values of 0.964, 0.846, and 0.842, respectively (Figure 5B). Differences among these groups were not significant (all P > 0.05) (Table 4). Optimal cutoff values for each subgroup, determined by the Bootstrap method, were 0.53, 0.61, and 0.51, respectively. The 95% confidence intervals for these values extensively overlapped (0.38–1.14, 0.59–1.37, 0.30–0.92) (Table 4).
TABLE 4
| Groups | AUC (95% CI) | Cut-off point | Bootstrap 95% CI for Cut-off | Sensitivity, % | Specificity, % | P-value (vs. TBI) | P-value (vs. ICH) | P-value (vs. SAH) | PPV, % | NPV, % | Youden index |
|---|---|---|---|---|---|---|---|---|---|---|---|
| TBI (n = 30) | 0.842 (0.663–0.948) | ≤ 0.53 | ≤0.38 to ≤ 1.14 | 69.23 | 100.00 | – | 0.9640 | 0.1740 | 100.0 | 81.0 | 0.69 |
| ICH (n = 47) | 0.846 (0.711–0.935) | ≤ 0.61 | ≤0.59 to ≤ 1.37 | 80.95 | 84.62 | 0.9640 | – | 0.1123 | 81.0 | 84.6 | 0.66 |
| SAH (n = 15) | 0.964 (0.725–1.000) | ≤ 0.51 | ≤0.3 to ≤ 0.92 | 85.71 | 100.00 | 0.1740 | 0.1123 | – | 100.0 | 88.9 | 0.86 |
Diagnostic performance of CGLR for intracranial infection: comparison across etiology subgroups.
CGLR, cerebrospinal fluid glucose-to-lactate ratio glucose-to-lactate ratio; TBI, traumatic brain injury; ICH, intracranial hemorrhage; SAH, subarachnoid hemorrhage.
3.3 Multivariable analyses
Independent predictors associated with intracranial infection were determined using multivariable logistic regression analysis. The model included variables significant in the univariate analysis (CGLR, CSF leukocyte, cPro, and CSF PCT) but excluded cGlu and cLac due to intrinsic collinearity with CGLR. In multivariate analysis, CGLR was the only independent predictor significantly associated with intracranial infection (OR: 0.112; 95% CI: 0.040–0.318; P < 0.001). In contrast, CSF PCT, CSF leukocyte count, and CSF protein did not retain independent significance.
4 Discussion
Glucose is a critical substrate for maintaining normal brain metabolism and function. In patients with postoperative intracranial infections after ABI, reduced brain glucose availability results from impaired cerebral blood flow, bacterial metabolism, neutrophil glycolysis, and anaerobic metabolism in brain tissue (, ). Recent studies have demonstrated rapid activation of microglial cells during inflammation or brain injury (, ). In this activated state, microglial cells have increased energy demands, leading to enhanced glucose uptake and metabolism. This metabolic shift is mediated by targets such as lactate dehydrogenase, mammalian target of rapamycin, and hypoxia-inducible factor-1 alpha, which increase glycolysis and lactate release, further exacerbating mitochondrial dysfunction (). This process may explain the elevated lactate and reduced glucose levels observed in CSF of ABI patients with postoperative infections. However, elevated CSF lactate can also occur without ischemia.
Previous studies indicated that CSF metabolites have high diagnostic accuracy in patients with suspected intracranial infections, particularly bacterial meningitis (, ). In this study, mean cGlu and cLac levels significantly differed between infected and non-infected groups. These findings align with studies by Peng et al. () and Pesaresi et al. (). Additionally, the areas under the ROC curve for cGlu and cLac were 0.836 and 0.820, respectively, demonstrating strong predictive capability for postoperative intracranial infection in ABI patients.
Limited studies have explored associations between biomarker concentrations in CSF and blood after ABI surgeries. No significant correlation emerged between bLac and cLac (r = 0.071, P = 0.502), consistent with previous findings (). This lack of correlation results from limited lactic acid permeability across the blood-brain and blood-CSF barriers, and the absence of active transport mechanisms for lactate (). Meanwhile, cGlu was significantly correlated with cLac (r = -0.56, P < 0.001), suggesting that elevated cLac relates to cerebral hyperglycolysis. However, cLac showed only a weak correlation with CSF PCT (r = 0.277, P = 0.009), and no significant correlation with cGlu (r = -0.100, P = 0.352). ROC analysis yielded an AUC of 0.631 for CSF PCT, consistent with previous results (, –).
Kul et al. () examined 56 TBI patients and reported that low CGLR combined with elevated CSF lactate indicated unfavorable outcomes. A retrospective study by Morotti and Goldstein (), including 114 patients with non-traumatic SAH across five European university intensive care units, identified an independent association between CGLR and poor prognosis. Similarly, a multicenter prospective observational study by Abassi et al. (), involving 219 ABI patients admitted to 12 European ICUs, concluded that decreased CGLR independently predicted adverse neurological outcomes at 3 months and ICU mortality. Recently, Wang et al. () identified CGLR as an early diagnostic biomarker for vasospasm in patients with aneurysmal SAH. Collectively, these studies support the view that CGLR is a versatile biomarker for assessing central nervous system metabolic status, consistent with our finding of strong diagnostic performance of CGLR across different ABI subtypes.
In this study, patients with ABI who developed intracranial infection after brain surgery had significantly lower CGLR values compared to non-infected individuals. The ROC curve analysis yielded an AUC of 0.866 for CGLR, demonstrating superior overall accuracy as a biomarker for intracranial infection compared to cLac or cGlu alone. At an optimal cutoff of 0.59, CGLR achieved a specificity of 94.1% and a sensitivity of 75.6%. Subgroup analyses confirmed consistent diagnostic performance of CGLR in patients with TBI, ICH, and SAH. The optimal cutoff values (0.51–0.61) for these subgroups had extensively overlapping 95% confidence intervals, statistically supporting a unified CGLR cutoff (0.59) for patients with different ABI types. From a clinical perspective, these findings suggest that a CGLR value ≤ 0.59, given its high specificity (94.1%), can help rule in intracranial infection with reasonable confidence, supporting the initiation or continuation of empirical antibiotic therapy. Conversely, a CGLR value > 0.59, with its moderate sensitivity (75.6%), may assist in ruling out infection in patients with low clinical suspicion, potentially allowing deferral or de-escalation of antibiotic therapy while awaiting culture results.
Kul et al. () previously reported that a CGLR below 1.47 independently predicted poor prognosis in ABI patients. Abassi et al. () identified an optimal CGLR cutoff of 1.39 to predict adverse outcomes at 3 months. The discrepancy in optimal cutoff values between studies may be due to additional metabolic disturbances caused by infection (, ), leading to rapid glucose depletion and a significant increase in lactate. Consequently, the cutoff (0.59) for diagnosing infection is lower, suggesting that CGLR sensitively reflects different pathological conditions. To address whether CGLR primarily reflects injury severity rather than infection, we compared CGLR between infected and non-infected patients. CGLR remained significantly lower in infected patients regardless of glycemic status, suggesting an infection-specific component beyond injury severity.
This study confirmed a modest positive correlation between peripheral and CSF glucose (r = 0.267, P = 0.001), consistent with physiological expectations. However, blood glucose levels did not differ significantly between infected and non-infected groups. Subgroup analysis revealed that CGLR remained significantly lower in infected patients, irrespective of glycemic status (all P < 0.001). No significant differences in CGLR appeared between normoglycemic and hyperglycemic patients within the same infection status (P = 0.932 for infected; P = 0.550 for non-infected). Additionally, a strong inverse correlation between CSF glucose and lactate was observed in infected patients (r = -0.549, P < 0.001), but not in non-infected individuals (r = -0.259, P = 0.069), suggesting infection induces a tightly coupled local metabolic state Zhang et al. (). Soon et al. recently reported excellent diagnostic accuracy of CLGR (AUC 0.994) in neurosurgical patients. However, their cohort had fewer culture-negative infections (8.5%) and excluded many inconclusive cases. In contrast, our cohort included more culture-negative infections (58.5%) and a broader spectrum of ABI etiologies, possibly explaining the relatively lower performance of CLGR in our population (AUC 0.832). In this study, CGLR slightly outperformed CLGR (AUC 0.866 vs. 0.832). A potential explanation is that CGLR, which does not directly depend on peripheral glucose levels, may be less affected by systemic glycemic variations. In contrast, CLGR incorporates the CSF-to-blood glucose ratio and may be more sensitive to stress-induced hyperglycemia or insulin therapy in critically ill ABI patients. Therefore, compared with CLGR, CGLR not only showed higher diagnostic accuracy (AUC 0.866 vs. 0.832) but also eliminates the need for simultaneous blood glucose measurement, reducing clinical inconvenience and potential timing-related errors. Moreover, it is relatively insensitive to peripheral glucose fluctuations and specifically reflects central nervous system infection status, offering distinct advantages over CSF culture (days) or NGS ( > 24 h, high cost).
The current study has several limitations. First, its findings are constrained by a limited sample size and single-center design, requiring multicenter validation to confirm the proposed CGLR cutoff. Second, dynamic analysis of CSF metabolites was not performed, preventing evaluation of the relationship between changes in intracranial infection severity and CSF metabolites. Third, despite controlling for multiple confounding factors, unmeasured variables may remain. Furthermore, since all patients were managed in the ICU due to severe ABI, the findings might not generalize to patients with mild ABI who do not require ICU care. Future studies are needed to validate CGLR in non-ICU settings. Finally, subtle variations in CSF glucose and lactate measurement methods among laboratories may challenge the establishment of a universal cutoff value.
5 Conclusion
Although using cLac or cGlu alone demonstrates good diagnostic efficacy for postoperative intracranial infection after ABI, the combined indicator CGLR shows superior diagnostic performance. Additionally, CGLR sensitively reflects various pathological conditions. Moreover, it is relatively insensitive to systemic glycemic fluctuations and more specifically reflects central nervous system infection status. Thus, CGLR may serve as a rapid, practical, and cost-effective biomarker for early diagnosis of postoperative intracranial infections in ABI patients.
Statements
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics statement
The studies involving humans were approved by the Medical Ethics Committee of the Affiliated People’s Hospital of Ningbo University The Affiliated People’s Hospital of Ningbo University. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.
Author contributions
WW: Writing – original draft, Writing – review & editing. HW: Supervision, Validation, Writing – review & editing. CZ: Conceptualization, Data curation, Writing – review & editing. YF: Conceptualization, Supervision, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was funded by the Medical Science and Technology Project of Zhejiang Province (2025KY1479) and Ningbo Clinical Research Center Member for Emergency and Critical Diseases under Grant (2024L003).
Acknowledgments
We gratefully acknowledge the invaluable contributions of our ICU team and the patients involved in this research.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1.
AbassiMBangdiwalaASNuwagiraEKandole TadeoKOkirwothMWilliamsDAet al. Cerebrospinal fluid lactate as a prognostic marker of disease severity and mortality in cryptococcal meningitis.Clin Infect Dis. (2021) 73:e3077–82. 10.1093/cid/ciaa1749
2.
American Diabetes Association2. Classification and diagnosis of diabetes: standards of medical care in Diabetes-2018.Diabetes Care. (2018) 41:S13–27. 10.2337/dc18-S002
3.
CowanMNSethiIHarrisTH. Microglia in CNS infections: insights from toxoplasma gondii and other pathogens.Trends Parasitol. (2022) 38:217–29. 10.1016/j.pt.2021.12.004
4.
de AlmeidaSMCastoldiJRRiechiSC. Comparison between cerebrospinal fluid biomarkers for differential diagnosis of acute meningitis.Diagnosis. (2023) 10:298–308. 10.1515/dx-2023-0013
5.
DjukicMLangePErbguthFNauR. Spatial and temporal variation of routine parameters: pitfalls in the cerebrospinal fluid analysis in central nervous system infections.J Neuroinflamm. (2022) 19:174. 10.1186/s12974-022-02538-3
6.
DuttaSSachdevaNPalARayP. Cerebrospinal fluid and plasma procalcitonin for the diagnosis of neonatal bacterial meningitis.J Paediatr Child Health. (2022) 58:1425–30. 10.1111/jpc.16023
7.
FangCZhuTZhangPXiaLSunC. Risk factors of neurosurgical site infection after craniotomy: a systematic review and meta-analysis.Am J Infect Control. (2017) 45:e123–34. 10.1016/j.ajic.2017.06.009.
8.
Gouvea BogossianETalebCAspideRBadenesRBattagliniDBilottaFet al. Cerebro-Spinal fluid glucose and lactate concentrations changes in response to therapies in patients with primary brain injury: the start-trip study.Crit Care. (2023) 27:130. 10.1186/s13054-023-04409-6
9.
HermanidesJHongYTTrivediMOuttrimJAigbirhioFNestorPJet al. Metabolic derangements are associated with impaired glucose delivery following traumatic brain injury.Brain. (2021) 144:3492–504. 10.1093/brain/awab255
10.
HuangQWangYChenSLiangF. Glycometabolic reprogramming of microglia in neurodegenerative diseases: insights from neuroinflammation.Aging Dis. (2024) 15:1155–75. 10.14336/AD.2023.0807
11.
KongYNiuJLiuZZhouWTianYShiG. Heparin-Binding protein as a diagnostic biomarker for healthcare-associated meningitis and ventriculitis in pediatric patients.Neurosurgery. (2025) 97:710–8. 10.1227/neu.0000000000003505
12.
KulGSencanIKulHKorkmazNAltunayE. The role of cerebrospinal fluid biomarkers in the diagnosis of post-neurosurgical meningitis.Turk Neurosurg. (2020) 30:513–9. 10.5137/1019-5149.JTN.25175-18.2
13.
LiRYangYWangHZhangTDuanFWuKet al. Lactate and lactylation in the brain: current progress and perspectives.Cell Mol Neurobiol. (2023) 43:2541–55. 10.1007/s10571-023-01335-7
14.
LozanoAFranchiFSeastresRJOddoMLheureuxOBadenesRet al. Glucose and lactate concentrations in cerebrospinal fluid after traumatic brain injury.J Neurosurg Anesthesiol. (2020) 32:162–9. 10.1097/ANA.0000000000000582
15.
MorottiAGoldsteinJN. Diagnosis and management of acute intracerebral hemorrhage.Emerg Med Clin North Am. (2016) 34:883–99. 10.1016/j.emc.2016.06.010
16.
NieJZhangWZhangHYuHLiALuoCet al. Development and validation of a predictive model for postoperative intracranial infections in neurosurgery with risk factor analysis.World Neurosurg. (2024) 189:e126–40. 10.1016/j.wneu.2024.05.184
17.
PanXHaishaerDLiuMZhouSNaHZhaoH. Diagnostic, monitoring, and prognostic value of combined detection of cerebrospinal fluid heparin-binding protein, Interleukin-6, Interleukin-10, and procalcitonin for post-neurosurgical intracranial infection.Cytokine. (2024) 179:156593. 10.1016/j.cyto.2024.156593
18.
PengQWangLYuCMChuXZhuBF. Diagnostic value of serum Nlrp3, Metalloproteinase-9 and interferon-gamma for postoperative hydrocephalus and intracranial infection in patients with severe craniocerebral trauma.Exp Physiol. (2024) 109:956–65. 10.1113/EP091463
19.
PesaresiABattagliniDAnaniaPSgambetterraSOrigliaCZonaGet al. Cerebrospinal fluid analysis and changes over time in patients with subarachnoid hemorrhage: a prospective observational study.J Anesth Analg Crit Care. (2025) 5:31. 10.1186/s44158-025-00250-1
20.
SharmaMSVohraAThomasPKapilASuriAChandraPSet al. Effect of risk-stratified, protocol-based perioperative chemoprophylaxis on nosocomial infection rates in a series of 31 927 Consecutive neurosurgical procedures (1994-2006).Neurosurgery. (2009) 64: 1123–30; discussion 30–1. 10.1227/01.NEU.0000345645.51847.61.
21.
SoonBJaafarASBakarAANarayananV. Combined lactate glucose ratio as a novel marker for rapid diagnosis of cerebrospinal fluid bacterial infection in neurosurgical patients: diagnostic accuracy study and benchtop analyzer correlation.World Neurosurg. (2024) 191: e607–21. 10.1016/j.wneu.2024.09.012
22.
StaalSLOlieSEvan WeeghelMSchomakersBVVazFMI-Pace Study Groupet al. Cerebrospinal fluid metabolome in central nervous system infections: a study of diagnostic accuracy.Ann Neurol. (2025) 98:851–63. 10.1002/ana.27291
23.
TacconeFSBadenesRAribSRubulottaFMirekSFranchiFet al. Cerebrospinal fluid glucose and lactate levels after subarachnoid hemorrhage: a multicenter retrospective study.J Neurosurg Anesthesiol. (2020) 32:170–6. 10.1097/ANA.0000000000000584
24.
WangHZhouCFuY. Factors influencing procalcitonin in the cerebrospinal fluid of patients after neurosurgery and its diagnostic value for intracranial infection.BMC Neurol. (2023) 23:288. 10.1186/s12883-023-03339-8
25.
WangQWangYYangYKongYPengY. The value of elevated cerebrospinal fluid lactate concentrations in post-neurosurgical bacterial meningitis.BMC Neurol. (2023) 23:377. 10.1186/s12883-023-03428-8
26.
YeYChenJXuJLuoQFuPZhaoFet al. The diagnostic and prognostic value of heparin-binding protein in cerebrospinal fluid for patients with intracranial infections.Eur J Med Res. (2024) 29:579. 10.1186/s40001-024-02183-x
27.
ZhangSLachanceBBMattsonMPJiaX. Glucose metabolic crosstalk and regulation in brain function and diseases.Prog Neurobiol. (2021) 204:102089. 10.1016/j.pneurobio.2021.102089
Summary
Keywords
acute brain injury, cerebrospinal fluid glucose, cerebrospinal fluid glucose-to-lactate, cerebrospinal fluid lactate, postoperative intracranial infection
Citation
Wang W, Wang H, Zhou C and Fu Y (2026) Cerebrospinal fluid glucose-to-lactate ratio (CGLR) as a diagnostic biomarker for postoperative intracranial infections in patients with acute brain injury: a prospective diagnostic accuracy study. Front. Med. 13:1830485. doi: 10.3389/fmed.2026.1830485
Received
14 March 2026
Revised
08 May 2026
Accepted
21 May 2026
Published
10 June 2026
Volume
13 - 2026
Edited by
Jian-Xin Zhou, Capital Medical University, China
Reviewed by
Bee Hong Soon, National University of Malaysia, Malaysia
Alessandro Pesaresi, Istituti Ospitalieri di Cremona, Italy
Updates
Copyright
© 2026 Wang, Wang, Zhou and Fu.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Huajun Wang, whj269696@163.com
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.