Cerebrospinal Fluid Pharmacokinetics of Tigecycline in Lumbar and Ventricular Compartments After Lumbar Intrathecal Administration: A Three-Patient Case Series

Abstract

Background

Lumbar intrathecal administration of tigecycline has been increasingly used as a salvage treatment for multidrug-resistant (MDR) intracranial infections. However, whether lumbar administration can generate adequate tigecycline exposure in ventricular cerebrospinal fluid (CSF), remains insufficiently characterized.

Case Presentation

We retrospectively analyzed three patients with intracranial infections who received 5mg tigecycline through a lumbar drainage catheter and had concurrent lumbar and ventricular CSF access. The underlying conditions were postoperative primary meningioma, brain abscess, and traumatic cerebral contusions, respectively. The identified pathogens included Acinetobacter johnsonii (tigecycline MIC, 0.5 mg/L), Aspergillus fumigatus, and Klebsiella pneumoniae (tigecycline MIC, 2.0 mg/L). Paired lumbar and ventricular CSF samples were collected at scheduled time points after dosing, and tigecycline concentrations were quantified by liquid chromatography–tandem mass spectrometry (LC‑MS/MS). Following a 5 mg lumbar intrathecal dose, the lumbar CSF peak concentration (C_max) was reached within 1 hour across all patients (933.89, 269.07, and 256.71 mg/L). In contrast, the ventricular C_max values were substantially lower (0.67, 7.05, and 2.64 mg/L), with ventricular time to peak concentration (T_max) delayed to 4–6 hours. The ventricular‑to‑lumbar area‑under‑the‑curve ratio (V/L AUC_1-24h ratio), ranged from 0.33% to 7.13%, indicating markedly attenuated and variable ventricular exposure relative to lumbar exposure. In the two patients with bacterial infections and available tigecycline MICs, ventricular CSF concentrations exceeded the corresponding MIC for at least part of the dosing interval. No neurotoxic adverse events were observed.

Conclusion

In this three-patient case series, 5 mg lumbar intrathecal tigecycline achieved measurable but markedly attenuated ventricular CSF exposure, with substantial interpatient variability, and no predefined neurotoxic adverse events were documented in this small series. Lumbar intrathecal administration may be considered when ventricular administration is unavailable or undesirable, with individualized assessment and therapeutic drug monitoring where feasible.

Introduction

Central nervous system (CNS) infections, including bacterial meningitis, brain abscesses, and ventriculitis, are life-threatening complications in neurosurgical intensive care. The increasing prevalence of MDR, extensively drug-resistant (XDR) and carbapenem-resistant Gram-negative bacteria has made antimicrobial treatment increasingly difficult,¹ particularly after trauma, neurosurgery, CSF leakage, or the placement of external ventricular drainage devices (EVDs).² Although systemic antimicrobial therapy remains the foundation of treatment, adequate CNS exposure is often difficult to achieve because many antibiotics have limited penetration across the blood-brain barrier (BBB) and the blood-CSF barrier. Delayed microbiological control is associated with neurological deterioration, prolonged hospitalization, and high mortality, highlighting the urgent need for effective strategies against MDR intracranial infections.

Intrathecal administration—delivered either via the intraventricular route or the lumbar route—has emerged as a valuable strategy that bypasses the BBB and achieves high local drug concentrations at the site of infection.^3,4 Current guidelines for healthcare-associated ventriculitis and meningitis recommend considering intraventricular therapy when infection responds poorly to systemic therapy alone, and suggest that intraventricular dosing be adjusted according to CSF antimicrobial concentrations, ventricular size, drainage volume, and the target pathogen MIC.⁴ Tigecycline is a potent broad-spectrum glycylcycline antibiotic with activity against many MDR Gram-positive and Gram-negative organisms, including carbapenem-resistant Acinetobacter baumannii and Klebsiella pneumoniae.^5,6 However, its clinical utility in CNS infections is extremely limited by poor drug penetration of the BBB after intravenous administration. Reported CSF exposure after systemic dosing is generally low, even in the presence of meningeal inflammation, and may be insufficient for organisms with higher MICs.^7,8 Consequently, although formal recommendations regarding the intrathecal use, optimal dose, and therapeutic targets of tigecycline remain unavailable, intrathecal or intraventricular tigecycline has emerged as a potential salvage option in selected patients with difficult-to-treat CNS infections. Published case reports and small series^2,9–13 have described favorable microbiological and clinical responses; however, pharmacokinetic information regarding its distribution within different CSF compartments remains limited.

Nevertheless, most published experience has focused on intraventricular administration or on clinical outcomes without paired compartmental pharmacokinetic sampling. Direct ventricular injection is pharmacologically attractive because it delivers the drug to the ventricular CSF,³ but it requires an Ommaya reservoir or ventricular catheter and may not be feasible or desirable in all patients. Lumbar intrathecal administration is technically simpler and less invasive in many centers, yet cephalad movement from the lumbar subarachnoid space to the ventricles must occur against the usual bulk flow of CSF. This raises an important clinical question: after lumbar intrathecal tigecycline, are ventricular concentrations adequate, or is exposure limited to the lumbar compartment?

To address this question, we reviewed three patients with intracranial infection who had simultaneous lumbar and ventricular CSF access during treatment with lumbar intrathecal tigecycline. By measuring tigecycline concentrations in both compartments, this case series aimed to characterize the lumbar-to-ventricular concentration gradient, quantify relative ventricular-to-lumbar exposure, and explore the relationship between ventricular exposure and pathogen MIC after a 5 mg lumbar intrathecal dose.

Case Presentation and Pharmacokinetic Sampling

We retrospectively reviewed the medical records of three patients with intracranial infections who received lumbar intrathecal tigecycline at Nanjing Drum Tower Hospital. All patients had both a lumbar drainage catheter and ventricular access in place during treatment, allowing paired lumbar and ventricular CSF sampling. The study was approved by the Institutional Ethics Committee of Nanjing Drum Tower Hospital. (Approval number: 2026-CR0008-01).

In all cases, tigecycline (5 mg, diluted in 5 mL of 0.9% sterile saline) was administered via the lumbar drainage catheter by slow manual injection. Following the injection, the lumbar drain was flushed with 5 mL of saline immediately to ensure complete drug delivery, and the lumbar catheter was clamped for 1 hour to reduce immediate drug loss through drainage and facilitate cephalad drug distribution. Paired CSF samples (2mL each) were simultaneously collected from both the lumbar drainage and the ventricular access (VP shunt or Ommaya) at 1, 2, 4, 6, 12, and 24 hours after dosing. For lumbar samples, the first 3mL of fluid aspirated from the catheter was discarded to clear dead space, while for ventricular samples, the initial 1 mL of stagnant tubing fluid was discarded before sampling. All CSF samples were immediately centrifuged at 3,000 rpm for 10 minutes at 4°C, and the supernatant was stored at −80°C until analysis.

Tigecycline concentrations in CSF were quantified using liquid chromatography–tandem mass spectrometry (LC–MS/MS) at the Clinical Department of Laboratory Medicine, Nanjing Drum Tower Hospital. The analytical platform had undergone performance verification at our institution. The assay showed linearity over a calibration range of 16–1600 ng/mL (R² = 0.9992), with a lower limit of quantification (LLOQ) of 16 ng/mL. Across low, medium, and high quality-control levels, intra-assay coefficients of variation (CVs) were ≤5.60% and inter-assay CVs were ≤4.91%, while mean recoveries ranged from 98.10% to 102.95%. Samples with concentrations above the upper limit of the calibration range were diluted before reanalysis, and the final concentrations were corrected for the corresponding dilution factor. Pharmacokinetic parameters included the maximum observed concentration (C_max), time to maximum observed concentration (T_max), area under the observed concentration–time curve from 1 to 24 h (AUC_1–24h), and the ventricular-to-lumbar AUC_1–24h ratio (V/L AUC_1–24h ratio). C_max and T_max were obtained directly from the observed concentration–time data. Concentration–time curves were generated using GraphPad Prism 10.1, and AUC_1–24h was calculated using the trapezoidal rule implemented in the software. The V/L AUC_1–24h ratio was calculated as ventricular AUC_1–24h divided by lumbar AUC_1–24h × 100% and was used as an exploratory measure of relative ventricular-to-lumbar exposure. Tigecycline MICs for the bacterial isolates were determined by the broth microdilution method using Mueller–Hinton broth in the clinical microbiology laboratory. Clinical data, including Glasgow Coma Scale (GCS) scores, daily CSF drainage volume, CSF biochemical parameters, adverse events, and outcomes, were extracted from the medical records.

Case 1

A 73-year-old woman underwent resection of a meningioma and subsequently developed intracranial infection. CSF culture grew Acinetobacter johnsonii, with a tigecycline MIC of 0.5 mg/L. At the pharmacokinetic study day, CSF protein concentration was 1704.8 mg/L, and the daily lumbar drainage volume was 300mL. At discharge, the Glasgow Coma Scale (GCS) score was 12.

Case 2

A 60-year-old woman with a brain abscess ultimately diagnosed as Aspergillus fumigatus infection received tigecycline as empirical antibacterial coverage before the definitive fungal diagnosis was established. CSF protein was 803.7 mg/L and daily lumbar drainage was 150mL. The discharge GCS score was 15.

Case 3

A 48-year-old man with traumatic cerebral contusions developed intracranial infection caused by Klebsiella pneumoniae, with a tigecycline MIC of 2.0 mg/L. CSF protein was 868.5 mg/L and daily lumbar drainage was 200mL. The discharge GCS score was 12.

Result

Baseline Characteristics

The demographic, microbiological, CSF, and drainage characteristics of the three patients are summarized in Table 1. The median age was 60 years (range 48–73 years). All patients had inflammatory CSF profiles at the time of pharmacokinetic sampling. Patient 1 had the highest CSF protein concentration and the greatest daily lumbar drainage volume, whereas patient 3 had a higher tigecycline MIC among the two patients with bacterial infections.

Table 1.

Baseline Clinical and Biochemical Characteristics of the Study Patients

Characteristics	Patient 1	Patient 2	Patient 3
Demographics
Age (years)	73	60	48
Sex (male/female)	Female	Female	Male
Body Mass Index (kg/m2)	27.06	19.53	26.12
Clinical status
Primary Diagnosis	Meningioma	Brain abscess	Traumatic cerebral contusions
GCS Score at Enrollment	15(E4V5M6)	15(E4V5M6)	8(E3V1M4)
Comorbidities	Glaucoma, Graves’ disease	None	Diabetes, Hypertension
Target pathogen (CSF)	Acinetobacter johnsonii	Aspergillus fumigatus	Klebsiella pneumoniae
Resistance phenotype	Ceftazidime- and fluoroquinolone-resistant; meropenem intermediate	N/A	Carbapenem-resistant K. pneumoniae (CRKP)
Selected antimicrobial susceptibility	IPM S; MEM I; CAZ R; FEP R; CIP R; LVX R; CST S	N/A	IPM R; MEM R; CAZ R; FEP R; CIP R; LVX R; CST S; CZA S
Tigecycline MIC (mg/L)	0.5	N/A	2.0
Concomitant IV Tigecycline (Yes/No)	No	No	No
Initial surgery	Brain tumor resection surgery	Cranial abscess incision and drainage surgery	Cerebral hematoma removal surgery
CSF parameters (at PK study)
WBC Count (10^6/L)	108	710	19
Protein (mg/L)	1704.8	803.7	868.5
Glucose (mmol/L)	4.44	1.76	3.23
Chloride (mmol/L)	131.6	113.0	123.9
Lactic acid (mmol/L)	2.3	3.6	2.1
Drainage Devices
Ventricular Access	VP shunt	Ommaya reservoir	Ommaya reservoir
Lumbar Access	Lumbar drainage	Lumbar drainage	Lumbar drainage
Daily CSF Drainage (mL/24h)	300	150	200
Systemic Laboratory Data (at PK study)
Serum Albumin (g/L)	44.3	33.4	32.2
Serum Creatinine (μmol/L)	37	62	38
CrCl (mL/min)	122.31	67.04	236.78
Outcome
Total Length of stay(days)	87	50	126
GCS at Discharge	12(E3V4M5)	15(E4V5M6)	12(E4V3M5)

Abbreviations: GCS, Glasgow Coma Scale; IV, intravenous; PK, pharmacokinetic; WBC, white blood cell; MIC, minimum inhibitory concentration; VP, ventriculoperitoneal; CrCl, creatinine clearance. S, susceptible; I, intermediate; R, resistant; IPM, imipenem; MEM, meropenem; CAZ, ceftazidime; FEP, cefepime; CIP, ciprofloxacin; LVX, levofloxacin; TGC, tigecycline; CST, colistin; CZA, ceftazidime/avibactam.

Lumbar and Ventricular Pharmacokinetics

Following 5 mg lumbar intrathecal tigecycline, a marked spatial concentration gradient was observed between the lumbar and ventricular CSF (Table 2). As illustrated by the individual concentration–time profiles in Figure 1, lumbar CSF C_max values were reached rapidly within 1 hour in all three patients, with individual values of 933.89, 269.07, and 256.71 mg/L. By contrast, ventricular C_max values were substantially lower and delayed, reaching 0.67, 7.05 and 2.64 mg/L, respectively, with a time to T_max in ventricular CSF ranging from 4–6 hours (Table 2). The V/L AUC_1-24h ratio, used here as an exploratory measure of relative ventricular-to-lumbar exposure, ranged from 0.33% to 7.13%.

Table 2.

Pharmacokinetic Parameters of Tigecycline in Lumbar and Ventricular CSF

Patient	Site	Cmax (mg/L)	Tmax (h)	AUC1-24h (mg·h/L)	V/L AUC1-24h Ratio (%)
Patient 1	Lumbar	933.89	1	3998	0.33
	Ventricular	0.67	6	13.08
Patient 2	Lumbar	269.07	1	712.1	7.13
	Ventricular	7.05	4	50.74
Patient 3	Lumbar	256.71	1	1411	2.15
	Ventricular	2.64	4	30.36

Figure 1.

Pharmacokinetic profiles of tigecycline in lumbar and ventricular CSF. Individual concentration-time curves are shown for Patient 1–3. In each panel, the left Y-axis (black circles) represents tigecycline concentrations in the lumbar CSF, and the right Y-axis (red squares) represents concentrations in the ventricular CSF. Shaded areas correspond to the area under the observed concentration-time curve from 1 to 24 h (AUC_1–24h) for each compartment. The horizontal dotted line indicates the measured pathogen MIC where available and is shown for descriptive comparison only. The figure illustrates the marked difference in tigecycline exposure between the lumbar and ventricular CSF compartments and the delayed ventricular T_max.

In the two patients with documented bacterial infections and available tigecycline MICs, ventricular CSF concentrations exceeded the respective pathogen MICs for several hours after the 5 mg dose. These concentration-to-MIC comparisons were considered exploratory because validated CSF PK/PD targets for intrathecal tigecycline have not been established. In patient 1, ventricular concentrations remained above the A. johnsonii MIC of 0.5 mg/L for approximately 20 hours, although the margin above the MIC was narrow. In patient 3, ventricular concentrations exceeded the K. pneumoniae MIC of 2.0 mg/L for approximately 4 hours. Patient 2 had a fungal infection (Aspergillus fumigatus) for which tigecycline has no established therapeutic role or MIC-based interpretive breakpoint; therefore, the pharmacokinetic data from this patient were used to describe distribution and safety rather than antibacterial target attainment. Despite the heterogeneous PK profiles, all three patients were discharged alive, with Glasgow Coma Scale (GCS) scores ranging from 12 to 15 at discharge.

Safety and Tolerability

No predefined neurotoxic adverse events were documented during the observation period following lumbar intrathecal administration of 5 mg tigecycline in the three patients. Vital sign monitoring and systemic laboratory parameters revealed no significant drug-related toxicities (Table 3). All patients exhibited neurological stability or improvement, with discharge GCS score ranging from 12 to 15.

Table 3.

Safety and Adverse Events of Intrathecal Tigecycline

Patient	Dose (mg)	Neurotoxicity (Seizures/Encephalopathy)	Chemical Meningitis	Systemic Adverse Events (Nausea/Vomiting)	Other Drug-Related AEs
1	5	None	None	None	None
2	5	None	None	None	None
3	5	None	None	None	None

Discussion

This retrospective case series provides, to our knowledge, the first simultaneous characterization of tigecycline concentrations in the lumbar and ventricular CSF following lumbar intrathecal administration, thereby enabling a paired quantitative comparison of tigecycline exposure in the lumbar and ventricular CSF compartments. The main observation was a marked lumbar-to-ventricular concentration gradient. Although tigecycline was detectable in ventricular CSF after lumbar administration, ventricular exposure was substantially lower and more variable than lumbar exposure, with V/L AUC_1-24h ratio ranging from 0.33% to 7.13%. These findings demonstrate measurable ventricular exposure after lumbar administration; however, relative ventricular exposure was markedly attenuated and showed substantial interpatient variability.

The observed lumbar-to-ventricular gradient is consistent with CSF physiology. CSF is mainly produced in the ventricles and flows caudally through the subarachnoid space before absorption into the venous system. Therefore, drugs injected into the lumbar subarachnoid space must migrate cephalad against the usual direction of CSF bulk flow to reach the ventricles. This process may be affected by inflammation, CSF obstruction, drainage volume, catheter location, and clamping protocols. Compared with published intraventricular tigecycline pharmacokinetic data, the ventricular exposure in our lumbar series was lower. Huang et al reported that 5 mg intraventricular tigecycline achieved a ventricular C_max of approximately 37.89 mg/L in a patient with XDR A. baumannii intracranial infection.¹³ In contrast, the ventricular C_max values after lumbar dosing in our patients were 0.67 to 7.05 mg/L. Accordingly, direct intraventricular administration may be preferable when ventriculitis or ventricular infection is the dominant therapeutic target, especially for pathogens with higher MICs. Nevertheless, lumbar intrathecal administration may remain useful when ventricular access is unavailable, when an additional ventricular puncture is considered high-risk, or when infection involves the spinal or basal cisternal CSF spaces.

The interpatient variability observed in this series is clinically relevant. Patient 1 had both the highest CSF protein concentration and the greatest daily lumbar drainage volume and showed the lowest V/L AUC_1–24h ratio. Elevated CSF protein may have contributed to the relatively low ventricular exposure; however, the underlying mechanism remains uncertain. Tigecycline protein binding is complex, concentration-dependent, and variable across experimental conditions,^14,15 and unbound tigecycline concentrations were not measured in the present study. Therefore, increased binding or sequestration of tigecycline in protein-rich CSF represents only one possible explanation and cannot be established from these data. High lumbar CSF drainage may also have reduced drug availability for ventricular distribution despite temporary catheter clamping. Although these mechanisms cannot be proven from three cases, they are biologically plausible and align with pharmacometric considerations for intrathecal antibiotics.³ Future studies should prospectively evaluate CSF flow characteristics, drainage protocols, catheter position, and both total and unbound tigecycline concentrations.

The two bacterial cases illustrate the potential importance of pathogen MIC when interpreting ventricular exposure. In Patient 1, ventricular concentrations remained above the measured A. johnsonii MIC of 0.5 mg/L for much of the dosing interval, whereas in Patient 3, concentrations exceeded the measured K. pneumoniae MIC of 2.0 mg/L for a shorter period. However, validated CSF pharmacokinetic/pharmacodynamic (PK/PD) targets for intrathecal tigecycline have not been established, and the relationship between CSF drug exposure, pathogen MIC, and clinical or microbiological response remains uncertain. Therefore, the concentration-to-MIC comparisons in this case series should be interpreted as exploratory observations rather than evidence of established PK/PD target attainment. This consideration is particularly important for Acinetobacter spp., for which established clinical tigecycline susceptibility breakpoints are lacking. Accordingly, the observed concentrations exceeding the measured MIC should not, by themselves, be interpreted as demonstrating adequate therapeutic ventricular exposure. Nevertheless, these findings suggest that pathogen susceptibility may be one of several factors to consider when interpreting ventricular exposure, together with infection location, CSF drainage conditions, route of administration, and therapeutic drug monitoring.

Patient 2 was ultimately diagnosed with a fungal infection and was therefore not included in the evaluation of antibacterial target attainment. Tigecycline was administered before the definitive fungal diagnosis had been established, when bacterial intracranial infection was still strongly suspected. At that time, the patient had shown an inadequate clinical response to empirical vancomycin plus meropenem therapy, raising concern for possible MDR bacterial involvement. Given the severity of the intracranial infection and the poor response to conventional empirical antibiotics, intrathecal tigecycline was considered a clinically reasonable empirical option. After the fungal etiology was confirmed, this case was retained only for pharmacokinetic and safety observations. The paired CSF samples from this patient nevertheless provided useful information on lumbar-to-ventricular tigecycline distribution after lumbar intrathecal administration.

Regarding safety, no obvious neurotoxic adverse events—including seizures, myoclonus, drug-induced encephalopathy, or chemical meningitis—were observed after lumbar intrathecal tigecycline in these three patients, despite lumbar CSF concentrations that far exceeded typical systemic peak levels. However, given the very small sample size and the observational nature of this case series, these findings should be interpreted descriptively and cannot establish the safety of lumbar intrathecal tigecycline. Published reports have described variable intrathecal and intraventricular tigecycline regimens, commonly 1 to 10 mg per dose, and neurotoxicity remains a concern, particularly with higher local exposure or repeated administration.^2,9–11,13 A previous meta-analysis of adjunctive intrathecal antibiotics for Gram-negative nosocomial meningitis (23 studies, 229 patients) reported local complications including chemical meningitis and seizures.¹⁶

Limitations

Several limitations should be acknowledged. First, this was a retrospective case series involving only three patients, and the substantial heterogeneity in underlying disease, causative pathogens, CSF drainage conditions, ventricular access, and inflammatory status limits the generalizability of the findings. Second, the concentration-time profiles were based on a limited number of sampling points and were therefore intended primarily for descriptive pharmacokinetic characterization rather than formal population pharmacokinetic modeling. Third, several physiological and treatment-related factors that may influence lumbar-to-ventricular drug distribution, including CSF flow dynamics, anatomical obstruction, catheter position, and drainage conditions, could not be systematically controlled. Fourth, unbound tigecycline concentrations were not measured, precluding direct assessment of the potential influence of CSF protein binding on drug distribution. Finally, validated CSF PK/PD targets for intrathecal tigecycline have not been established, and the very small sample size precludes conclusions regarding clinical efficacy or safety. These findings should therefore be considered exploratory and hypothesis-generating.

Conclusion

In this three-patient case series, 5 mg lumbar intrathecal tigecycline achieved measurable but markedly attenuated ventricular CSF exposure, with substantial interpatient variability, and no predefined neurotoxic adverse events were documented in this small series. Lumbar intrathecal administration may be considered when ventricular administration is unavailable or undesirable, with individualized assessment and therapeutic drug monitoring where feasible.

Funding Statement

This study was supported by grants from the Beijing Medical Award Foundation (YXJL-2022-0351-0421), the Beijing Bethune Charitable Foundation (2022-YJ-085-J-Z-ZZ-026 and 2024-YJ-156-J-003), and the Nanjing Drum Tower Hospital Clinical Research Fund Project (2024-LCYJ-MS-08).

Ethics Approval and Consent to Participate

This study was approved by the Institutional Ethics Committee of Nanjing Drum Tower Hospital (approval number: 2026-CR0008-01). No additional institutional approval was required for publication of the de-identified case details beyond the ethics approval and written informed consent for publication. Written informed consent for publication of the clinical details and any accompanying de-identified data was obtained from the patient or the patients’ legally authorized representatives, as appropriate. Patient 2 had a Glasgow Coma Scale score of 15 and retained full decision-making capacity; written informed consent for publication was therefore obtained directly from this patient. Patients 1 and 3 lacked full decision-making capacity at the time consent was obtained, and written informed consent for publication was therefore obtained from their legally authorized representatives. The manuscript content intended for publication was reviewed by the patients or their legally authorized representatives before consent was provided. All patient information was de-identified, and de-identification did not alter the scholarly meaning of the clinical information presented.

Disclosure

The authors declare that they have no competing interests in this work.