Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2026 Jun 9;21(6):e0351306. doi: 10.1371/journal.pone.0351306

Early mortality risk stratification in childhood bacterial meningitis using cerebrospinal fluid glucose and protein levels and their combinations: A multicontinental cohort study

Markku Kallio 1,*,#, Tuula Pelkonen 1,2,3,#, Irmeli Roine 4,#, Heikki Peltola 1,#
Editor: Nilanka Perera5
PMCID: PMC13249170  PMID: 42263098

Abstract

Background

Bacterial meningitis remains a major cause of death and neurological disability in children worldwide, particularly in low-resource settings where access to intensive care is limited. Cerebrospinal fluid (CSF) glucose, protein, and leukocyte concentrations are routinely measured at hospital presentation, yet their potential for early risk stratification – individually or in combination – has not been clearly defined. Because mortality risk is often difficult to assess at admission, identifying simple CSF-based thresholds that help clinicians recognize high-risk patients could improve triage and management globally.

Methods and findings

We evaluated CSF glucose, protein and leukocyte levels in 1598 children with bacteriologically confirmed bacterial meningitis across six countries in Latin America, Angola, and Finland to determine their individual and combined associations with mortality and neurological outcome. CSF glucose, protein, and leukocyte counts were measured at admission and examined in relation to in-hospital mortality and neurological outcome at discharge, as assessed by the Glasgow Outcome Scale. CSF glucose and protein concentrations were categorized into five clinically interpretable groups. Mortality increased progressively with decreasing CSF glucose (from 8% to 29%) and with increasing protein concentrations (from 5% to 23%). Children with both low CSF glucose (<10 mg/dL) and high CSF protein (≥200 mg/dL) had a mortality rate of 27% compared with 1% among those with normal CSF glucose and protein. A CSF leukocyte count <500/µL was also associated with increased mortality. Lower CSF glucose (ρ = 0.24; P < .001) and higher protein (ρ = 0.15; P < .001) were both associated with poorer neurological outcomes.

Conclusions

Routine CSF markers available at the time of hospital admission provide prognostic information that supports risk stratification in pediatric bacterial meningitis. Simple CSF-based thresholds—particularly when CSF glucose and protein are interpreted in combination—may help identify high-risk children early and guide clinical prioritization in settings with limited critical care capacity.

Introduction

Bacterial meningitis remains a severe and life-threatening disease in children, with substantial mortality and neurological sequelae despite global vaccination programs and advances in antimicrobial therapy [1,2]. The burden is greatest in low- and middle-income countries, where delays in diagnosis and limited access to intensive care contribute to high case fatality rates. Early risk assessment is crucial for guiding management and resource allocation, yet the prognostic information embedded in routine cerebrospinal fluid (CSF) measurements – particularly glucose and protein – has not been fully utilized in clinical practice. Clinical presentation of bacterial meningitis is variable: although the classical triad of fever, headache, and vomiting is well recognized, many children present with nonspecific findings, making early diagnosis challenging [3–6]. Consequently, lumbar puncture remains essential for early evaluation in all suspected cases.

Because the causative pathogen is seldom identified at admission except by Gram stain, clinicians rely on routine CSF analyses to guide diagnosis and early treatment decisions. In bacterial meningitis, CSF glucose is typically reduced, protein is elevated, and leukocyte counts are increased, reflecting inflammatory processes within the central nervous system [7,8]. However, laboratory findings vary widely between patients. Approximately 6% of children with culture-confirmed bacterial meningitis have normal CSF leukocyte counts, and lymphocyte predominance is observed in up to 10% of cases [9]. Such variability limits the diagnostic and prognostic value of CSF leukocyte counts alone.

CSF glucose decreases due to blood-brain barrier disruption, increased glycolytic activity of inflammatory cells, and bacterial consumption of glucose [6,8]. In contrast, elevated CSF protein reflects increased blood-brain barrier permeability in severe inflammation and has been associated with poor outcomes [10–12]. Although CSF glucose and protein are not direct inflammatory biomarkers, they are universally available, low-cost indicators of central nervous system involvement – even in settings lacking advanced laboratory capacity. Several prognostic models have been developed to improve risk stratification in meningitis [13–17], including the Bacterial Meningitis Score [18] and the Meningitis Score for Emergencies [19]. However, these tools are designed primarily to distinguish bacterial from viral meningitis rather than to predict mortality, and their applicability in resource-limited settings is restricted by dependence on laboratory parameters that are not routinely available. Importantly, CSF glucose – long recognized as a marker of poor prognosis – is not explicitly incorporated into commonly used prognostic tools [5].

Over the years, our group has conducted five large prospective studies of childhood bacterial meningitis across three continents – six countries in Latin America, Angola, and Finland – yielding a clinically diverse dataset that reflects real-world conditions in both low- and high-resource settings. This dataset enables evaluation of two key clinical questions with direct global relevance: [1] Can mortality risk in childhood bacterial meningitis be predicted using routine CSF measurements at admission to hospital, based on clinically interpretable thresholds for glucose, protein, and leukocytes? and [2] Do combinations of these CSF markers identify children at highest risk of death?

Methods

Patients and data collection

We conducted five prospective bacterial meningitis trials in between 1984 and 2017 in Finland [20], 6 countries in Latin America [21], and Angola [22], with detailed descriptions published elsewhere [23]. Number of patients in Finland (n = 334), in Angola (n = 721), and in Latin American countries (n = 543): Argentina (n = 131), Brazil (n = 14), Dominican Republic (n = 106), Ecuador (n = 112), Venezuela (n = 108), Paraguay (n = 72). The data used in this study were combined into a single dataset on 15/9/2020. Bacterial meningitis was confirmed if any of the following criteria were met: positive CSF culture; compatible clinical features with a positive blood culture; positive CSF latex agglutination test; positive CSF Gram stain; or positive CSF polymerase chain reaction [20–23]. Exclusion criteria included age younger than 2 months, trauma, intracranial shunt, preexisting hearing impairment or neurologic disease, and use of immunosuppressive medication. Patients diagnosed with encephalitis or tuberculous meningitis were excluded from the study. On admission, the attending pediatrician performed a physical examination and lumbar puncture, ordered relevant tests, and completed a standardized questionnaire used across all trials. The child’s presenting condition was graded with the age-adjusted Glasgow Coma Scale (range 3–15) [24].

At hospital discharge, a clinical evaluation was performed with special attention to neurologic and audiologic outcomes. Outcomes were assessed using the modified Glasgow Outcome Score (range, 1–5), with 1 indicating uneventful recovery and 5 indicating death. Intermediate categories included survival with mild hearing deficit (better-ear threshold, 41–79 dB), deafness (≥80 dB), and mild or severe neurologic sequelae. Mild sequelae included hemiparesis, monoparesis, psychomotor retardation, or ataxia, whereas severe sequelae included blindness, quadriplegia/paresis, hydrocephalus requiring a shunt, or severe psychomotor retardation.

CSF and blood glucose concentrations were measured, CSF protein levels assessed, CSF leukocytes counted, and bacteria identified at each institution using standard methods. Normal reference values were defined as CSF glucose concentration ≥40 mg/dL, CSF protein concentration <60 mg/dL, CSF leukocytes <5/mm³, and a CSF-to-blood glucose ratio of 0.4. The large number of bacteriologically confirmed meningitis cases (n = 1598) allowed analysis of CSF variables by dividing concentrations into 5 clinically practical and easily interpretable groups, ensuring sufficient numbers in each category for trend analyses. The objective was to determine whether simple, clinically interpretable thresholds could be identified to help clinicians assess mortality risk. CSF glucose was categorized as normal (≥40 mg/dL), 30 to <40 mg/dL, 20 to <30 mg/dL, 10 to <20 mg/dL, and <10 mg/dL. CSF protein was categorized as <60 mg/dL, 60–99 mg/dL, 100–199 mg/dL, 200 to <299 mg/dL, and ≥300 mg/dL. CSF leukocytes were categorized as <100/mm³, 100–499/mm³, 500–999/mm³, 1000–1999/mm³, and ≥2000/mm³. Data from daily completed questionnaires were collected and analyzed in the original studies, and all data were subsequently combined and analyzed in Finland [20–23].

Ethics approval and consent to participate

The Ethics Committee of Luanda Children’s Hospital approved the studies conducted in Angola, and the hospital boards approved the studies in Latin America and Finland (1984–2017). Following the initiation of international clinical trial registration, the 2 Angolan studies were registered (ISRCTN62824827; April 10, 2005, and NCT01540838; February 29, 2012). Patients were enrolled only after written or oral informed consent was obtained from their legal guardian; for illiterate guardians, a fingerprint served as consent. All methods were conducted in accordance with the Declaration of Helsinki.

Statistical analysis

Data analysis was performed using JMP Pro 15.0 (SAS Institute Inc) for MacOS. Descriptive data are presented as counts with percentages or medians with interquartile ranges (IQRs), as appropriate. Differences in baseline characteristics between groups were assessed using the Wilcoxon or Kruskal-Wallis tests, χ² tests, or the Mann-Whitney test, depending on data type. Associations between ordinal categorical exposures and binary outcomes were evaluated with the Cochran-Armitage trend test, which detects trends across ordered groups in a binary response variable. Glasgow Outcome Score results were compared with CSF glucose, protein, and leukocyte counts (continuous data) using Spearman correlation. Nominal logistic regression analysis was used to evaluate survival outcomes, with odds ratios and 95% CIs calculated for predictor variables. The Effect Likelihood Ratio test was applied to the final model, demonstrating significant contributions from CSF glucose, CSF protein and altered consciousness. Model performance was evaluated using a Receiver Operating Characteristic (ROC) curve, with the Area Under the Curve (AUC) indicating a moderate ability to differentiate between survival and mortality.

Results

A total of 721 children from Angola, 543 from Latin America, and 334 from Finland comprised the series of 1598 cases of bacterial meningitis caused by Haemophilus influenzae (n = 668), Streptococcus pneumoniae (n = 506), Neisseria meningitidis (n = 285), or other bacteria (n = 139). The etiology was determined by CSF culture in 1225 cases (77%), CSF latex agglutination test in 97 (6%), CSF Gram stain in 124 (8%), CSF polymerase chain reaction in 98 (6%), and blood culture in 54 (3%). Males accounted for 862 cases (55%) and females for 696 cases (45%). The median age was 1.2 years (IQR, 0.6–3.2) for boys and 1.0 years (IQR, 0.5–2.8) for girls. Country-specific laboratory and clinical variables are presented in Table 1.

Table 1. Patient characteristics and cerebrospinal fluid (CSF) values at admission and during illness. Data are presented as median values; percentages or interquartile ranges (25th - 75th percentile) are shown in parentheses, as appropriate.

Whole series Finland Latin America Angola
At arrival
Age – years 1.1 (0.5-3.1) 1.8 (0.8-3.4) 0.8 (0.5-2.6) 1.0 (0.5-2.9)
Ill before arrival – days 3 (2 –5) 1 (1 –3) 2 (2–3) 4 (3 –7)
Pretreatment antibiotics (%) 502/1454 (35%) 58/323 (18%) 169/494 (34%) 275/637 (43%)
Decreased level of consciousness
(Glasgow coma score <15) (%)		 1112/1473 (75%)
 263/325 (79%) 401/512 (78%) 448/636 (72%)
Seizures before / at arrival (%) 585/1515 (39%) 60/328 (18%) 182/513 (35%) 343/674 (51%)
Axillary temperature °C 38.0 (37.0-39.0) 39.0 (38.3-39.7) 38.0 (37.0-38.6) 37.7 (37.0-38.5)
CRP (mg/L) 158 (91-161) 130 (78-195) 152 (73-169) 161 (110-161)
Blood leucocytes (103/µL) 14.8 (9.1-21.0) 13.6 (8.5-20.0) 14.9 (9.0-21.1) 15.3 (9.8-22.1)
CSF Glucose (mg/dL)
Survived 16 (7–37)a 32 (13-60) 14 (5–33)a 13 (6–24)b
Died 10 (5 –17) 7 (2-69) 6 (3 –15) 10 (6 –17)
CSF Protein (mg/dL)
Survived 160 (90–250)a 160 (90–240)b 150 (90–250)b 190 (120–260)c
Died 230 (170-370) 390 (190-480) 250 (150-470) 210 (180-280)
CSF Leucocytes (/mm3)
Survived 2010 (700–5700)c 3030 (1270–6900)c 1960 (680–6200)c 1500 (430–3500)c
Died 1330 (320-4470) 2400 (480-8300) 1560 (400-6830) 1250 (290-3790)
During illness course
Days in hospital 10 (8 –12) 11 (9 –11) 8 (7 –11) 11 (9 –17)
Deafness d (%) 110/1034 (11%) 22/271 (8%) 45/375 (12%) 43/388 (11%)
Neurological sequelae (%) 380/1194 (32%) 56/322 (17%) 133/450 (29%) 191/422 (45%)
Open in a new tab

a = p < 0.001 b = p < 0.01 c = p < 0.05 These differences are between survived and died patients

d = better ear's hearing threshold ≥80 dB

CSF glucose and survival

Lower CSF glucose concentrations were associated with higher mortality (Fig 1). A significant linear trend was observed between decreasing CSF glucose levels and increased mortality (Cochran-Armitage trend test, P < .001; χ² test, P < .001). For overall outcomes at discharge, measured by the Glasgow Outcome Score, CSF glucose as a continuous variable correlated with outcome (Spearman ρ = 0.24; P < .001).

Fig 1. Survival vs death in children with bacterial meningitis by cerebrospinal fluid (CSF) glucose concentration at hospital admission.

Fig 1

Open in a new tab

Odds ratios compare normal glucose values (≥40 mg/dL) with lower concentrations categorized into 4 groups. a = p < 0.001 b = p < 0.01 c = p < 0.05.

Because the study period was long, during which vaccines were introduced and treatment practices changed, we compared outcomes across each decade. The prognostic value of low CSF glucose and high CSF protein for mortality remained consistent across all time periods. Similarly, the findings were unchanged when analyses were stratified by the major bacterial pathogens or patients with and without antibiotic pretreatment.

CSF protein and survival

Higher CSF protein concentrations were significantly associated with increased mortality (Fig 2). A significant linear trend was observed across protein categories (Cochran-Armitage trend test, P < .001; χ² test, P < .001). For overall outcomes at discharge, measured by the Glasgow Outcome Score, CSF protein as a continuous variable correlated with outcome (Spearman ρ = 0.16; P < .001).

Fig 2. Survival vs death in children with bacterial meningitis by cerebrospinal fluid (CSF) protein concentration at hospital admission.

Fig 2

Open in a new tab

Odds ratios compare normal protein values (<60 mg/dL) with higher concentrations, categorized into 4 groups. a = p<0.001 c = p<0.05.

CSF leucocyte and survival

CSF leukocyte count, one of the most commonly used diagnostic indices in bacterial meningitis, was categorized into five groups. A low CSF leukocyte count was associated with higher mortality. Children admitted with counts <500/mm³ had approximately a 50% greater risk of death compared with those with higher counts. However, once the leukocyte count exceeded 500/mm³, no additional prognostic value was observed, whether the count was 1000/mm³, 2000/mm³, or higher (Cochran-Armitage trend test, P = non-significant). For overall outcomes at discharge, measured by the Glasgow Outcome Score, CSF leukocyte count as a continuous variable correlated with outcome (Spearman ρ = 0.13; P < .01).

The CSF-to-blood glucose ratio and survival

The CSF-to-blood glucose ratio was less strongly associated with mortality than CSF glucose concentration alone, although the trend remained significant (Cochran-Armitage trend test, P < .01; χ² test, P < .05). When the ratio was only moderately decreased (≥0.20), the trend was less evident.

CSF glucose and survival in Streptococcus pneumoniae meningitis

Because the incidence of Haemophilus influenzae and Neisseria meningitidis meningitis has markedly decreased with vaccination, we analyzed the association between CSF glucose levels and mortality in Streptococcus pneumoniae meningitis. Mortality was significantly higher at lower CSF glucose levels compared with higher levels (Fig 3). A significant linear trend was observed between decreasing CSF glucose concentrations and increased mortality (Cochran-Armitage trend test, P < .001; χ² test, P < .001).

Fig 3. Survival vs death in children with Streptococcus pneumoniae meningitis by cerebrospinal fluid (CSF) glucose concentration at hospital admission.

Fig 3

Open in a new tab

Odds ratios compare normal CSF glucose values (≥40 mg/dL) with lower concentrations, categorized into 4 groups. a = p < 0.001.

Survival in patients with Markedly Low CSF Glucose and High CSF Protein

Because both reduced CSF glucose and elevated CSF protein were strongly associated with mortality, we evaluated their combined prognostic effect by comparing children with markedly abnormal values to those with both parameters within the normal range (Fig 4). Mortality was 27% (46 of 169) among patients with concomitantly low CSF glucose <10 mg/dL and high CSF protein >200 mg/dL values, compared with 1% (1 of 90) among patients with normal CSF glucose and CSF protein values (χ² = 35.1; P < .001; Odds ratio, 31.7; 95% CI, 4.4–237.0).

Fig 4. Survival vs death in children with bacterial meningitis comparing those with normal cerebrospinal fluid (CSF) glucose (>40 mg/dL) and CSF protein (<60 mg/dL) values with those with very low CSF glucose (<10 mg/dL) and very high CSF protein (>200 mg/dL) values.

Fig 4

Open in a new tab

Odds ratio compares these 2 groups. a = p < 0.001.

The survival of patients with decreased level of consciousness (Glasgow coma scale under 15) and its association with varying CSF glucose or CSF protein levels

A Glasgow coma scale under 15 upon hospital admission correlates with a poorer prognosis in bacterial meningitis. In our dataset, approximately 75% of patients presented to the emergency department with a GCS < 15, enabling us to study how CSF glucose and protein levels behaved in these patients in relation to survival. Patients with a GCS < 15, but normal or near to normal CSF glucose (>30 mg/dL) and protein levels (<100 mg/dL) had a better prognosis compared to the other groups. Conversely, if patients with a GCS < 15, had low CSF glucose (<10 mg/dL) or high CSF protein (>200 mg/dL), survival was significantly poorer.

Predictors of survival in bacterial meningitis

To investigate predictors of survival in bacterial meningitis, a nominal logistic regression analysis was conducted using CSF glucose, CSF protein, altered consciousness, age, sex, pulse, pretreatment antibiotics, study region and duration of illness at hospital admission as explanatory variables. Age, sex, pulse, pretreatment antibiotics, study region and duration of illness were initially included in the model but were found to be non-significant (p > 0.05) and were subsequently excluded from the final analysis. In the final model, the Effect Likelihood Ratio -test indicated significant contributions from CSF glucose (Chi Square = 23.7, p < 0.001), CSF protein (Chi Square = 22.1, p < 0.001) and altered consciousness (GCS < 15) (Chi Square = 14.2, p < 0.001). The model’s discriminatory ability was assessed using the Receiver Operating Characteristic (ROC) curve, which yielded an Area Under the Curve (AUC) of 0.74, indicating a moderate ability to distinguish between survival and mortality outcomes. The ROC analysis was based on the final logistic regression model, which included the significant predictors identified in the Effect Likelihood Ratio test: CSF glucose, CSF protein, altered consciousness.

Discussion

CSF analysis is central to the diagnosis and management of bacterial meningitis, yet its potential contribution to early prognostic assessment remains underexplored. In this multicontinental cohort of 1598 children with bacteriologically confirmed meningitis, we found a clear linear association between decreasing CSF glucose, increasing CSF protein, and mortality. These findings are consistent with earlier observations linking low CSF glucose to adverse outcomes [6,10,25,26], but our study extends previous work by demonstrating that mortality risk increases progressively across glucose and protein strata rather than at a single critical threshold. Importantly, children with combined low CSF glucose and high CSF protein were at markedly increased risk of death, suggesting that simultaneous impairment of glucose transport and blood-brain barrier integrity reflects severe meningeal inflammation and metabolic distress. Because CSF glucose and protein are among the first laboratory results available at admission – even in resource-limited settings – these findings have direct clinical relevance, providing simple parameters that may support early risk stratification and guide triage decisions before etiological confirmation. The associations between these CSF abnormalities and poorer neurological outcomes at discharge further highlight their value as early indicators of disease severity.

Glucose is the primary metabolic substrate for the brain and is transported into the cerebrospinal fluid through facilitated diffusion across the blood-brain barrier. In bacterial meningitis, the inflammatory response disrupts blood-brain barrier function and impairs glucose transport, while activated leukocytes and bacteria increase glycolytic consumption of glucose within the subarachnoid space [27]. These mechanisms explain the progressive decline in CSF glucose observed in our study and its strong association with mortality. We found that severely low CSF glucose levels, particularly below 10 mg/dL, were associated with markedly increased risk of death. The flattening of the risk gradient beyond these thresholds (<10 mg/dL and <20 mg/dL) suggests a point of metabolic failure, at which compensatory mechanisms are insufficient to maintain cerebral energy supply. Low CSF glucose has previously been associated not only with mortality but also with neurological complications such as sensorineural hearing loss [28], further supporting its role as a marker of severe disease.

Although measurement of the cerebrospinal fluid (CSF)-to-blood glucose ratio is traditionally recommended to aid in the diagnosis of bacterial meningitis [8,29,30]. In our cohort, the ratio was a less sensitive predictor of survival than CSF glucose or protein concentrations alone. In routine practice, its use is constrained by practical limitations: blood and CSF samples must be obtained simultaneously, which is often not feasible in emergency settings, particularly after the initiation of intravenous fluids. In many of our patients, blood glucose values were influenced by glucose-containing maintenance fluids that were initiated when the child presented to the emergency department, which distorted the (CSF)-to-blood glucose ratio and reduced its discriminative value. Furthermore, when CSF glucose levels are severely reduced, the ratio remains low regardless of blood glucose concentration, reducing the clinical relevance of the denominator.

Under physiological conditions, the CSF contains only small amounts of plasma-derived proteins that cross the blood-brain barrier through selective filtration. In bacterial meningitis, inflammatory injury disrupts endothelial junctions and increases barrier permeability, allowing excess serum proteins to enter the subarachnoid space. Consistent with previous reports linking elevated CSF protein to adverse outcomes [10,17,31,32], we found that the risk of death increased progressively with rising protein concentrations. Markedly elevated CSF protein levels (>200 mg/dL), particularly those exceeding 300 mg/dL, were associated with substantially worse prognosis compared with only mildly elevated values. This pattern likely reflects the extent of blood-brain barrier disruption and the severity of the inflammatory response within the central nervous system.

The combination of CSF glucose and protein provided substantially greater prognostic discrimination than either marker alone. Mortality in bacterial meningitis was low among children with both markers within the normal range, whereas approximately one-third of those with severely low CSF glucose (<10 mg/dL) and markedly elevated protein (>300 mg/dL) died. These findings suggest that concurrent failure of glucose transport and significant blood-brain barrier disruption reflect severe inflammatory pathology and high early mortality risk. In contrast, although CSF leukocyte count remains essential for diagnosis, its prognostic value was limited. Leukocyte counts below 500/mm³ were associated with increased mortality, consistent with impaired host immune response, but values above this threshold did not differentiate survivors from non-survivors, indicating a ceiling effect. In severe infections such as meningitis with septic shock, CSF leukocyte counts may be unexpectedly low, reflecting overwhelming infection and impaired host response. Low CSF leukocyte counts have been shown to be independently associated with adverse outcomes in bacterial meningitis [6,33]

In clinical practice, mortality risk in childhood bacterial meningitis is often difficult to assess at the time of admission, particularly in settings where clinical severity scores with advanced laboratory testing are unavailable. Our findings show that simple combinations of routine CSF measurements – glucose and protein – can identify children at high risk of death using thresholds that are immediately interpretable by clinicians. These results support incorporation of CSF glucose and protein combinations into early risk assessment and triage protocols and may provide a foundation for pragmatic severity classification in future clinical trials. Because these measurements are already part of standard diagnostic evaluation worldwide, they could be implemented rapidly without additional cost, training, or equipment. Further studies are needed to evaluate whether CSF-based thresholds can guide intensified monitoring or adjunctive therapy and to determine their prognostic value in culture-negative disease. Tools that leverage existing diagnostics rather than requiring new technologies have potential for global impact, particularly in low-resource settings where most deaths from bacterial meningitis occur and timely triage remains a challenge.

This study has several strengths. It includes one of the largest prospectively collected cohorts of childhood bacterial meningitis with bacteriological confirmation and standardized data collection across three continents. The multinational design captures a broad clinical spectrum from both high-resource and low-resource settings, increasing the external validity of the findings. Importantly, the prognostic assessment was based on routine CSF parameters available at admission, enhancing clinical applicability even in hospitals with limited diagnostic resources.

This study also has limitations. Differences in socioeconomic and clinical conditions across the 3 continents may have influenced outcomes. We attempted to mitigate the long study duration by focusing on admission CSF findings, which were independent of treatment protocols. Patient care in all hospitals included supportive measures such as oxygen and fluid therapy, and laboratory practices for CSF analysis were consistent across sites and over time. Focusing on bacteriologically confirmed cases enhances reliability but may constrain generalizability to children with suspected meningitis and abnormal CSF findings without confirmed bacterial etiology. The prognostic value of CSF glucose and protein in culture-negative cases remains uncertain, highlighting the need for further research in this population, particularly in resource-limited settings where culture results may be unavailable or delayed.

Conclusion

This multicontinental analysis shows that simple cerebrospinal fluid measurements obtained at admission can stratify mortality risk in childhood bacterial meningitis without reliance on advanced diagnostics. These findings support the development of pragmatic, severity-based triage strategies using existing laboratory resources. Validation of these thresholds in prospective studies, including children with culture-negative meningitis, will be essential to guide future clinical application and to reduce preventable deaths in high-burden regions.

We confirm that:

• All authors have made substantial contributions to the work,

• All authors have approved the final manuscript,

• The manuscript has not been published previously and is not under consideration by any other journal,

• There are no conflicts of interest to declare

Data Availability

The data contain potentially identifiable patient information and are therefore not publicly available. De-identified data may be shared with qualified researchers upon reasonable request and ethics committee approval by contacting the Helsinki University Hospital (HUS) ethics committee at (eettinen.toimikunta@hus.fi).

Funding Statement

The author(s) received no specific funding for this work.

References

  • 1.Mohanty S, Johansson Kostenniemi U, Silfverdal SA, Salomonsson S, Iovino F, Bencina G, et al. Adult Work Ability Following Diagnosis of Bacterial Meningitis in Childhood. JAMA Netw Open. 2024;7(12):e2445497. doi: 10.1001/jamanetworkopen.2024.45497 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Liu Y, Feng Y, Guo Y, Chen J, Liu C, Liang J. Clinical predictors of poor outcome of bacterial meningitis in infants less than 90 days: a systematic review. Front Pediatr. 2024;12:1414778. doi: 10.3389/fped.2024.1414778 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Curtis S, Stobart K, Vandermeer B, Simel DL, Klassen T. Clinical features suggestive of meningitis in children: a systematic review of prospective data. Pediatrics. 2010;126(5):952–60. doi: 10.1542/peds.2010-0277 [DOI] [PubMed] [Google Scholar]
  • 4.Brouwer MC, Thwaites GE, Tunkel AR, van de Beek D. Dilemmas in the diagnosis of acute community-acquired bacterial meningitis. Lancet. 2012;380(9854):1684–92. doi: 10.1016/S0140-6736(12)61185-4 [DOI] [PubMed] [Google Scholar]
  • 5.Hasbun R. Progress and Challenges in Bacterial Meningitis: A Review. JAMA. 2022;328(21):2147–54. doi: 10.1001/jama.2022.20521 [DOI] [PubMed] [Google Scholar]
  • 6.van de Beek D, de Gans J, Spanjaard L, Weisfelt M, Reitsma JB, Vermeulen M. Clinical features and prognostic factors in adults with bacterial meningitis. N Engl J Med. 2004;351(18):1849–59. doi: 10.1056/NEJMoa040845 [DOI] [PubMed] [Google Scholar]
  • 7.Fitch MT, van de Beek D. Emergency diagnosis and treatment of adult meningitis. Lancet Infect Dis. 2007;7:191–200. [DOI] [PubMed] [Google Scholar]
  • 8.Spanos A, Harrell FE Jr, Durack DT. Differential diagnosis of acute meningitis. An analysis of the predictive value of initial observations. JAMA. 1989;262(19):2700–7. [PubMed] [Google Scholar]
  • 9.Tunkel AR, Hartman BJ, Kaplan SL, Kaufman BA, Roos KL, Scheld WM, et al. Practice guidelines for the management of bacterial meningitis. Clin Infect Dis. 2004;39(9):1267–84. doi: 10.1086/425368 [DOI] [PubMed] [Google Scholar]
  • 10.Tan J, Kan J, Qiu G, Zhao D, Ren F, Luo Z, et al. Clinical Prognosis in Neonatal Bacterial Meningitis: The Role of Cerebrospinal Fluid Protein. PLoS One. 2015;10(10):e0141620. doi: 10.1371/journal.pone.0141620 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Klinger G, Chin CN, Beyene J, Perlman M. Predicting the outcome of neonatal bacterial meningitis. Pediatrics. 2000;106(3):477–82. doi: 10.1542/peds.106.3.477 [DOI] [PubMed] [Google Scholar]
  • 12.Pelkonen T, Kallio M, Latvala T, Roine I, Peltola H. Clinical blindness in conjunction with childhood bacterial meningitis. Sci Rep. 2023;13(1):15530. doi: 10.1038/s41598-023-41685-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lovera D, Amarilla S, Araya S, Galeano F, González N, Martínez de Cuellar C, et al. Risk Factors for Death and Severe Neurological Sequelae in Childhood Bacterial Meningitis. Pediatr Emerg Care. 2022;38(12):637–43. doi: 10.1097/PEC.0000000000002651 [DOI] [PubMed] [Google Scholar]
  • 14.Olson D, Lamb MM, Gaensbauer JT, Todd JK, Halsey NA, Asturias EJ, et al. Risk Factors for Death and Major Morbidity in Guatemalan Children with Acute Bacterial Meningitis. Pediatr Infect Dis J. 2015;34(7):724–8. doi: 10.1097/INF.0000000000000720 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.de Jonge RCJ, van Furth AM, Wassenaar M, Gemke RJBJ, Terwee CB. Predicting sequelae and death after bacterial meningitis in childhood: a systematic review of prognostic studies. BMC Infect Dis. 2010;10:232. doi: 10.1186/1471-2334-10-232 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Wee LYJ, Tanugroho RR, Thoon KC, Chong CY, Choong CT, Krishnamoorthy S, et al. A 15-year retrospective analysis of prognostic factors in childhood bacterial meningitis. Acta Paediatr. 2016;105(1):e22-9. doi: 10.1111/apa.13228 [DOI] [PubMed] [Google Scholar]
  • 17.Mao D-H, Miao J-K, Zou X, Chen N, Yu L-C, Lai X, et al. Risk Factors in Predicting Prognosis of Neonatal Bacterial Meningitis-A Systematic Review. Front Neurol. 2018;9:929. doi: 10.3389/fneur.2018.00929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Nigrovic LE, Kuppermann N, Malley R. Development and validation of a multivariable predictive model to distinguish bacterial from aseptic meningitis in children in the post-Haemophilus influenzae era. Pediatrics. 2002;110(4):712–9. doi: 10.1542/peds.110.4.712 [DOI] [PubMed] [Google Scholar]
  • 19.Mintegi S, García S, Martín MJ, Durán I, Arana-Arri E, Fernandez CL, et al. Clinical prediction rule for distinguishing bacterial from aseptic meningitis. Pediatrics. 2020;146:e20201126. [DOI] [PubMed] [Google Scholar]
  • 20.Peltola H, Anttila M, Renkonen O-V; The Finnish Study Group. Randomised comparison of chloramphenicol, ampicillin, cefotaxime and ceftriaxone for childhood bacterial meningitis. Lancet. 1989;333:1281–7. [DOI] [PubMed] [Google Scholar]
  • 21.Peltola H, Roine I, Fernández J, Zavala I, Ayala SG, Mata AG, et al. Adjuvant glycerol and/or dexamethasone to improve the outcomes of childhood bacterial meningitis: a prospective, randomized, double-blind, placebo-controlled trial. Clin Infect Dis. 2007;45(10):1277–86. doi: 10.1086/522534 [DOI] [PubMed] [Google Scholar]
  • 22.Pelkonen T, Roine I, Cruzeiro ML, Pitkäranta A, Kataja M, Peltola H. Slow initial β-lactam infusion and oral paracetamol to treat childhood bacterial meningitis: a randomised, controlled trial. Lancet Infect Dis. 2011;11(8):613–21. doi: 10.1016/S1473-3099(11)70055-X [DOI] [PubMed] [Google Scholar]
  • 23.Peltola H, Roine I, Kallio M, Pelkonen T. Outcome of childhood bacterial meningitis on three continents. Sci Rep. 2021;11(1):21593. doi: 10.1038/s41598-021-01085-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kirkham FJ, Newton CRJC, Whitehouse W. Paediatric coma scales. Dev Med Child Neurol. 2008;50(4):267–74. doi: 10.1111/j.1469-8749.2008.02042.x [DOI] [PubMed] [Google Scholar]
  • 25.Glimåker M, Johansson B, Grindborg Ö, Bottai M, Lindquist L, Sjölin J. Adult bacterial meningitis: earlier treatment and improved outcome following guideline revision promoting prompt lumbar puncture. Clin Infect Dis. 2015;60(8):1162–9. doi: 10.1093/cid/civ011 [DOI] [PubMed] [Google Scholar]
  • 26.van de Beek D, de Gans J, Tunkel AR, Wijdicks EFM. Community-acquired bacterial meningitis in adults. N Engl J Med. 2006;354(1):44–53. doi: 10.1056/NEJMra052116 [DOI] [PubMed] [Google Scholar]
  • 27.Krishnan S, Chang AC, Stoltz BM, Prasadarao NV. Escherichia coli K1 Modulates Peroxisome Proliferator-Activated Receptor γ and Glucose Transporter 1 at the Blood-Brain Barrier in Neonatal Meningitis. J Infect Dis. 2016;214(7):1092–104. doi: 10.1093/infdis/jiw306 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Eisenhut M, Meehan T, Batchelor L. Cerebrospinal fluid glucose levels and sensorineural hearing loss in bacterial meningitis. Infection. 2003;31(4):247–50. doi: 10.1007/s15010-003-3208-0 [DOI] [PubMed] [Google Scholar]
  • 29.Chekrouni N, Kroon M, Drost EHGM, van Soest TM, Bijlsma MW, Brouwer MC, et al. Characteristics and prognostic factors of bacterial meningitis in the intensive care unit: a prospective nationwide cohort study. Ann Intensive Care. 2023;13(1):124. doi: 10.1186/s13613-023-01218-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Tamune H, Takeya H, Suzuki W, Tagashira Y, Kuki T, Honda H, et al. Cerebrospinal fluid/blood glucose ratio as an indicator for bacterial meningitis. Am J Emerg Med. 2014;32(3):263–6. doi: 10.1016/j.ajem.2013.11.030 [DOI] [PubMed] [Google Scholar]
  • 31.Weisfelt M, van de Beek D, Spanjaard L, Reitsma JB, de Gans J. Clinical features, complications, and outcome in adults with pneumococcal meningitis: a prospective case series. Lancet Neurol. 2006;5(2):123–9. doi: 10.1016/S1474-4422(05)70288-X [DOI] [PubMed] [Google Scholar]
  • 32.Koedel U, Scheld WM, Pfister H-W. Pathogenesis and pathophysiology of pneumococcal meningitis. Lancet Infect Dis. 2002;2(12):721–36. doi: 10.1016/s1473-3099(02)00450-4 [DOI] [PubMed] [Google Scholar]
  • 33.Bijlsma MW, Brouwer MC, Kasanmoentalib ES, Kloek AT, Lucas MJ, Tanck MW, et al. Community-acquired bacterial meningitis in adults in the Netherlands, 2006-14: a prospective cohort study. Lancet Infect Dis. 2016;16(3):339–47. [DOI] [PubMed] [Google Scholar]
PLoS One. 2026 Jun 9;21(6):e0351306. doi: 10.1371/journal.pone.0351306.r001

Author response to Decision Letter 0

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

25 Nov 2025

Decision Letter 0

Nilanka Perera

22 Mar 2026

-->PONE-D-25-60645-->-->Early mortality risk stratification in childhood bacterial meningitis using cerebrospinal fluid glucose and protein levels and their combinations: a multicontinental cohort study-->-->PLOS One

Dear Dr. kallio,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by May 06 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you’re ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

  • A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Nilanka Perera, MD, PhD

Academic Editor

PLOS One

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1.Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that there are restrictions to data sharing for this study. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

Before we proceed with your manuscript, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., a Research Ethics Committee or Institutional Review Board, etc.). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of recommended repositories, please see

https://journals.plos.org/plosone/s/recommended-repositories. You also have the option of uploading the data as Supporting Information files, but we would recommend depositing data directly to a data repository if possible.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. In the online submission form you indicate that your data is not available for proprietary reasons and have provided a contact point for accessing this data. Please note that your current contact point is a co-author on this manuscript. According to our Data Policy, the contact point must not be an author on the manuscript and must be an institutional contact, ideally not an individual. Please revise your data statement to a non-author institutional point of contact, such as a data access or ethics committee, and send this to us via return email. Please also include contact information for the third party organization, and please include the full citation of where the data can be found.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

5. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewer's Responses to Questions

-->Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

Reviewer #2: Partly

**********

-->2. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #2: No

**********

-->4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: This is a comprehensive and much-needed study with clear application to day-to-day practice.

Some minor revisions are suggested as below:

1. An indicator of the timing of the initial lumbar puncture and CSF analysis.

2. Clinically, the percentage drop in CSF glucose relative to the concurrent plasma glucose would be more helpful to the practicing clinician. Suggest a graph comparing the percentage of sugar drop with outcomes

Reviewer #2: I congratulate the authors for the excellent study which is clinically oriented and very useful for day to day practice. It is a large study with more than 1500 patients and is multi-centre. Here are some minor concerns which I would like to raise.

Description of Study Population (Latin America)

It would improve clarity to explicitly list the individual Latin American countries included in the study and the number of participants recruited from each country. This will help readers better understand the geographic distribution and heterogeneity of the cohort.

Very Long Study Period (1984–2017)

The study spans 33 years, during which substantial changes have occurred in:

Epidemiology of pathogens due to vaccination (Hib, pneumococcal vaccines)

Antibiotic regimens and resistance patterns

Supportive and intensive care management

Diagnostic modalities (PCR availability, improved microbiology techniques)

Pooling all data across this period may introduce temporal bias unless the authors demonstrate that pathogen distribution, management strategies, and outcomes were comparable across eras. A subgroup analysis by time period (e.g., pre- and post-vaccine eras) is strongly recommended.

Inclusion of Encephalitis Cases

Were patients with encephalitis included in this cohort? If so, please clarify the diagnostic criteria and describe their number and characteristics. If not, this should be clearly stated.

Inclusion of Tuberculous Meningitis

Were patients with tuberculous meningitis included? If yes, please provide details. If excluded, this should be explicitly mentioned in the methods.

Breakdown of the “Other Bacteria” Category

The “other” pathogen group is heterogeneous. Certain organisms (e.g., Listeria monocytogenes, Mycoplasma pneumoniae) may cause lymphocytic predominance with less pronounced CSF glucose reduction or protein elevation but can still result in severe outcomes.

A detailed breakdown of pathogens within the “other” category is essential. If numbers are small and heterogeneity substantial, consider:

A pathogen-specific sub-analysis

Or restricting conclusions to the major pathogens (H. influenzae, S. pneumoniae, N. meningitidis)

Description of Treatment Protocols

Given the long study period and multicontinental setting, a detailed description of antibiotic regimens, adjunctive therapies (e.g., dexamethasone), and supportive care strategies is warranted. Differences in management could significantly influence mortality and may confound the associations observed.

CSF Leukocyte Sub-analysis

The analysis includes total CSF leukocyte counts but does not explore:

Neutrophil vs lymphocyte predominance

Neutrophil/lymphocyte ratio

There is clinical evidence that very high CSF neutrophil counts may correlate with worse outcomes. A sub-analysis considering cell differentials and pathogen type would be informative, especially given that some bacteria produce lymphocytic responses and prior antibiotic therapy can alter CSF profiles.

Effect of Pretreatment Antibiotics

Approximately 30–35% of patients received antibiotics prior to CSF analysis. Antibiotic therapy is known to alter CSF glucose, protein, and cellular profiles.

It would be important to clarify whether:

The findings apply equally to pretreated and non-pretreated patients

A subgroup analysis was performed stratified by prior antibiotic exposure

Need for Multivariable Regression Analysis

Multiple clinical and demographic factors influence mortality in bacterial meningitis. The current analysis is largely unadjusted.

A multivariable logistic regression model adjusting for:

Age

Region

Pathogen

GCS

Seizures

Pretreatment antibiotics

Duration of illness

is necessary to determine whether CSF glucose and protein independently predict mortality.

Categorization of Continuous Variables

Several continuous variables were categorized into groups for analysis. While clinically intuitive, categorization reduces statistical power and may distort risk gradients.

Consider modeling CSF glucose and protein as continuous variables in regression analysis, potentially using spline modeling to assess non-linear effects.

Predictive Performance Metrics (ROC Curve)

The manuscript discusses “risk stratification,” yet no predictive performance metrics are presented.

Including:

ROC curves

Area under the curve (AUC)

Sensitivity and specificity at key thresholds

would strengthen claims of prognostic utility.

Low CSF Leukocyte Count and Mortality

The statement that “children admitted with counts <500/mm³ had approximately a 50% greater risk of death” appears inconsistent with common clinical observations.

Further exploration is warranted:

Were these patients immunosuppressed?

Did they present late?

Were specific pathogens overrepresented in this subgroup?

This finding requires deeper analysis and discussion.

Wide Confidence Interval in Combined Analysis (Figure 4)

The combined glucose/protein subgroup shows a very large odds ratio (OR 31.7) with a wide confidence interval (4.4–237), indicating limited precision and likely small sample size.

Absolute numbers should be clearly presented, and interpretation should be cautious.

Restriction to Bacteriologically Confirmed Cases

The study includes only confirmed cases. While this enhances diagnostic certainty, it limits applicability to real-world settings where many patients are culture-negative.

This limitation should be explicitly reflected in the conclusions, clarifying that findings apply only to confirmed bacterial meningitis.

Statement Regarding CSF-to-Blood Glucose Ratio

The statement suggesting practical limitations in measuring simultaneous blood glucose may be misleading. In routine practice, plasma glucose measurement at the time of lumbar puncture remains standard of care.

The findings of this study do not justify minimizing the importance of the CSF-to-blood glucose ratio in clinical practice, and this statement should be revised or removed.

**********

-->6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: No

Reviewer #2: Yes:Pramith Ruwanpathirana

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

Attachment

Submitted filename: PLOS ONE- review.docx

pone.0351306.s001.docx (22.9KB, docx)
PLoS One. 2026 Jun 9;21(6):e0351306. doi: 10.1371/journal.pone.0351306.r003

Author response to Decision Letter 1

29 Apr 2026

Dear Nilanka Perera,

Academic Editor, Plos One

Thank you for the opportunity to respond to the reviewers’ comments. Please find below our response letter, in which we have addressed the reviewers’ comments and implemented all requested revisions to the manuscript. We have also conducted the additional analyses suggested, where feasible.

In the final section of the letter, we provide our responses to the Editors’ comments as well.

Sincerely,

Markku Kallio, MD, PhD

Reviewer #1: This is a comprehensive and much-needed study with clear application to day-to-day practice.

Some minor revisions are suggested as below:

Reviewer’s Comment:

1. An indicator of the timing of the initial lumbar puncture and CSF analysis.

Our response: This was a large multicenter study conducted across six South American countries, Angola, and Finland, and precise timing data for the initial lumbar puncture and CSF analysis were unfortunately not recorded. However, across all participating hospitals, efforts were made to evaluate and manage patients with suspected meningitis as promptly as possible.

Reviewer’s Comment:

2. Clinically, the percentage drop in CSF glucose relative to the concurrent plasma glucose would be more helpful to the practicing clinician. Suggest a graph comparing the percentage of sugar drop with outcomes.

Our response: In many acutely ill children, glucose-containing intravenous fluids are initiated as soon as the child arrives at the emergency department. This can affect blood glucose measurements and may partly explain why the CSF-to-blood glucose ratio performed suboptimally in our dataset. In our view, a similar limitation would persist even when expressing the measure as the percentage decrease in CSF glucose relative to the concurrent plasma glucose. We revised our text of CSF glucose in Discussion.

Reviewer #2: I congratulate the authors for the excellent study which is clinically oriented and very useful for day to day practice. It is a large study with more than 1500 patients and is multi-centre. Here are some minor concerns which I would like to raise.

Reviewer’s Comment:

Description of Study Population (Latin America). It would improve clarity to explicitly list the individual Latin American countries included in the study and the number of participants recruited from each country. This will help readers better understand the geographic distribution and heterogeneity of the cohort.

Our response: The numbers of patients from each country have now been added to the manuscript in Methods

Reviewer’s Comment:

Very Long Study Period (1984–2017). The study spans 33 years, during which substantial changes have occurred in: Epidemiology of pathogens due to vaccination (Hib, pneumococcal vaccines), Antibiotic regimens and resistance patterns. Supportive and intensive care management, Diagnostic modalities (PCR availability, improved microbiology techniques).

Pooling all data across this period may introduce temporal bias unless the authors demonstrate that pathogen distribution, management strategies, and outcomes were comparable across eras. A subgroup analysis by time period (e.g., pre- and post-vaccine eras) is strongly recommended.

Our response: The following answer is added to Results section: Because the study period was long, during which vaccines were introduced and treatment practices changed, we compared outcomes across each decade. The prognostic value of low CSF glucose and high CSF protein for mortality remained consistent across all time periods. Similarly, the findings were unchanged when analyses were stratified by the major bacterial pathogens.

Reviewer’s Comment:

Inclusion of Encephalitis Cases. Were patients with encephalitis included in this cohort? If so, please clarify the diagnostic criteria and describe their number and characteristics. If not, this should be clearly stated.

Our response: Only bacterial meningitis were included. This is added in material and methods.

Reviewer’s Comment:

Inclusion of Tuberculous Meningitis. Were patients with tuberculous meningitis included? If yes, please provide details. If excluded, this should be explicitly mentioned in the methods.

Our response: They were excluded. This is added in material and methods.

Reviewer’s Comment:

Breakdown of the “Other Bacteria” Category. The “other” pathogen group is heterogeneous. Certain organisms (e.g., Listeria monocytogenes, Mycoplasma pneumoniae) may cause lymphocytic predominance with less pronounced CSF glucose reduction or protein elevation but can still result in severe outcomes.

A detailed breakdown of pathogens within the “other” category is essential. If numbers are small and heterogeneity substantial, consider A pathogen-specific sub-analysis. Or restricting conclusions to the major pathogens (H. influenzae, S. pneumoniae, N. meningitidis)

Our response: The ‘other bacteria’ group comprised 139 patients in total. The Cochran–Armitage test for trend remained statistically significant (p < 0.01). However, this group included 20 different bacterial species, resulting in very small numbers within each subgroup. Therefore, pathogen-specific trend analyses were not feasible due to insufficient statistical power

Reviewer’s Comment:

Description of Treatment Protocols. Given the long study period and multicontinental setting, a detailed description of antibiotic regimens, adjunctive therapies (e.g., dexamethasone), and supportive care strategies is warranted. Differences in management could significantly influence mortality and may confound the associations observed.

Our response: Given the long study period and multicontinental setting, treatment protocols varied across centers and over time, including differences in antibiotic regimens, use of adjunctive therapies (such as dexamethasone), and supportive care strategies. However, the prognostic value of low CSF glucose and high CSF protein for mortality remained consistent across these treatment variations. This is demonstrated in our aforementioned comparison between patients treated in different decades, and it is now reported in the Results section.

Reviewer’s Comment:

CSF Leukocyte Sub-analysis. The analysis includes total CSF leukocyte counts but does not explore: Neutrophil vs lymphocyte predominance, Neutrophil/lymphocyte ratio. There is clinical evidence that very high CSF neutrophil counts may correlate with worse outcomes. A sub-analysis considering cell differentials and pathogen type would be informative, especially given that some bacteria produce lymphocytic responses and prior antibiotic therapy can alter CSF profiles.

Our response: Due to the multicenter nature of the study, conducted primarily in Africa and Latin America, CSF neutrophil/lymphocyte analyses were not consistently available.

Reviewer’s Comment:

Effect of Pretreatment Antibiotics. Approximately 30–35% of patients received antibiotics prior to CSF analysis. Antibiotic therapy is known to alter CSF glucose, protein, and cellular profiles. It would be important to clarify whether the findings apply equally to pretreated and non-pretreated patients.

Our response: There were no differences between patients with and without antibiotic pretreatment; low CSF glucose and high CSF protein were equally significant predictors in both groups. This has been added to the Results section

Reviewer’s Comment:

Need for Multivariable Regression Analysis. Multiple clinical and demographic factors influence mortality in bacterial meningitis. The current analysis is largely unadjusted. A multivariable logistic regression model adjusting for: Age, Region, Pathogen, GCS, Seizures, Pretreatment antibiotics, Duration of illness is necessary to determine whether CSF glucose and protein independently predict mortality.

Our response: We have added nominal logistic Regression analysis in the Result section

Reviewer’s Comment:

Categorization of Continuous Variables. Several continuous variables were categorized into groups for analysis. While clinically intuitive, categorization reduces statistical power and may distort risk gradients. Consider modeling CSF glucose and protein as continuous variables in regression analysis, potentially using spline modeling to assess non-linear effects.

Our response: We agree that continuous-variable modeling is statistically important and have therefore included multivariable analyses using CSF glucose and CSF protein as continuous predictors. However, our clinical objective was also to identify pragmatic cutoff values for routine bedside use. In emergency care, clinicians frequently interpret CSF glucose and CSF protein categorically as “markedly abnormal” or not, yet evidence-based thresholds for particularly dangerous values are lacking. We therefore present threshold-based analyses as a clinically complementary approach intended to provide practical support for early risk assessment.

Reviewer’s Comment:

Predictive Performance Metrics (ROC Curve). The manuscript discusses “risk stratification,” yet no predictive performance metrics are presented.

Including: ROC curves, Area under the curve (AUC), Sensitivity and specificity at key thresholds, would strengthen claims of prognostic utility.

Our response: The discriminatory performance of the final model was evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) is reported in the Results section. The model demonstrated acceptable accuracy in identifying patients at increased risk of mortality, supporting its potential utility for risk stratification. Notably, the concurrent presence of markedly decreased cerebrospinal fluid (CSF) glucose and elevated CSF protein levels was strongly associated with increased mortality risk, likely reflecting a pronounced inflammatory response.

Reviewer’s Comment:

Low CSF Leukocyte Count and Mortality. The statement that “children admitted with counts <500/mm³ had approximately a 50% greater risk of death” appears inconsistent with common clinical observations. Further exploration is warranted:

Were these patients immunosuppressed? Did they present late? Were specific pathogens overrepresented in this subgroup? This finding requires deeper analysis and discussion.

Our response: In severe infections such as meningitis with septic shock, CSF leukocyte counts may be unexpectedly low, reflecting overwhelming infection and impaired host response. Low CSF leukocyte counts have been shown to be independently associated with adverse outcomes in bacterial meningitis (references number 6 and a new reference number 27). This text has been added to Discussion.

Reviewer’s Comment:

Wide Confidence Interval in Combined Analysis (Figure 4). The combined glucose/protein subgroup shows a very large odds ratio (OR 31.7) with a wide confidence interval (4.4–237), indicating limited precision and likely small sample size. Absolute numbers should be clearly presented, and interpretation should be cautious.

Our response: The wide confidence interval reflects the small number of events in the reference group. Only one of 89 children with normal CSF glucose and protein levels died, whereas 44 of 122 children with markedly low CSF glucose and markedly elevated CSF protein died. The absolute values are shown in the figure.

Reviewer’s Comment:

Restriction to Bacteriologically Confirmed Cases. The study includes only confirmed cases. While this enhances diagnostic certainty, it limits applicability to real-world settings where many patients are culture-negative. This limitation should be explicitly reflected in the conclusions, clarifying that findings apply only to confirmed bacterial meningitis.

Our response: We edited the following sentence in Discussion in the Study Limitations section: Focusing on bacteriologically confirmed cases enhances reliability but may constrain generalizability to children with suspected meningitis and abnormal CSF findings without confirmed bacterial etiology.

Reviewer’s Comment:

Statement Regarding CSF-to-Blood Glucose Ratio. The statement suggesting practical limitations in measuring simultaneous blood glucose may be misleading. In routine practice, plasma glucose measurement at the time of lumbar puncture remains standard of care. The findings of this study do not justify minimizing the importance of the CSF-to-blood glucose ratio in clinical practice, and this statement should be revised or removed.

Our response: The reviewer is correct in noting that plasma glucose measurement at the time of lumbar puncture remains the standard of care. Our intention was to highlight the practical observation that many critically ill pediatric patients with meningitis had already been started on glucose-containing intravenous fluids in the emergency department prior to lumbar puncture. Consequently, in some cases, blood glucose measurements may have provided somewhat unreliable values. This may, in certain situations, further lead to distortion of the CSF-to-blood glucose ratio, potentially yielding somewhat inaccurate values. We revised our text in Discussion.

Response to Academic Editor:

Academic Editor’s Comment: If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information:

Our response: The data underlying this study cannot be made publicly available due to ethical and legal restrictions. This study is a prospective, multicenter investigation conducted across eight countries on three continents, involving pediatric patients with meningitis. Ethical approval was obtained from the relevant Institutional Review Boards and Research Ethics Committees in each participating country and hospital prior to study initiation. The approved study protocols and informed consent procedures did not include provisions for unrestricted public data sharing in open-access repositories.

The dataset contains detailed clinical information on a vulnerable population (children with meningitis), collected across multiple international sites. Due to the combination of clinical variables, geographic information, and the relative rarity of certain presentations, there is a residual risk of re-identification. Furthermore, for participating sites within the European Union, data processing and sharing are subject to the General Data Protection Regulation (GDPR), which imposes strict requirements on the handling of personal health.

For these reasons, the ethics committees that approved the study have not granted permission for public deposition of the dataset. Data can be made available to qualified researchers upon reasonable request for scientifically valid purposes, subject to review and, where required, the relevant local ethics committees. Access will require a formal data use agreement to ensure compliance with applicable ethical, legal, and data protection requirements. Additional approvals from the originating site-specific ethics committees may be required prior to data sharing.

Academic Editor’s Comment: In the online submission form you indicate that your data is not available for proprietary reasons and have provided a contact point for accessing this data. Please note that your current contact point is a co-author on this manuscript. According to our Data Policy, the contact point must not be an author on the manuscript and must be an institutional contact, ideally not an individual. Please revise your data statement to a non-author institutional point of contact, such as a data access or ethics committee:

Our response: Here is the contact mail address to the Helsinki University Hospital (HUS) ethics committee

eettinen.toimikunta@hus.fi

Attachment

Submitted filename: Response to Reviewers.docx

pone.0351306.s003.docx (25.5KB, docx)

Decision Letter 1

Nilanka Perera

26 May 2026

<p>Early mortality risk stratification in childhood bacterial meningitis using cerebrospinal fluid glucose and protein levels and their combinations: a multicontinental cohort study

PONE-D-25-60645R1

Dear Dr. kallio,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Nilanka Perera, MD, PhD

Academic Editor

PLOS One

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.-->

Reviewer #1: All comments have been addressed

**********

-->2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

**********

-->3. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

**********

-->4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

**********

-->5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

**********

-->6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: No further comments. The authors have satisfactorily answered the comments raised by both reviewers. Statistical analysis of data is sound. Recommend acceptance for publication

**********

-->7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: No

**********

Acceptance letter

Nilanka Perera

PONE-D-25-60645R1

PLOS One

Dear Dr. kallio,

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS One. Congratulations! Your manuscript is now being handed over to our production team.

At this stage, our production department will prepare your paper for publication. This includes ensuring the following:

* All references, tables, and figures are properly cited

* All relevant supporting information is included in the manuscript submission,

* There are no issues that prevent the paper from being properly typeset

You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps.

Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing.

If we can help with anything else, please email us at customercare@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Nilanka Perera

Academic Editor

PLOS One

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Attachment

    Submitted filename: PLOS ONE- review.docx

    pone.0351306.s001.docx (22.9KB, docx)
    Attachment

    Submitted filename: Response to Reviewers.docx

    pone.0351306.s003.docx (25.5KB, docx)

    Data Availability Statement

    The data contain potentially identifiable patient information and are therefore not publicly available. De-identified data may be shared with qualified researchers upon reasonable request and ethics committee approval by contacting the Helsinki University Hospital (HUS) ethics committee at (eettinen.toimikunta@hus.fi).

    Articles from PLOS One are provided here courtesy of PLOS

    RESOURCES