Intrathecal (IT) drug delivery, which involves injecting drugs directly into the subarachnoid space, has become an indispensable technique in neuroscience research. By circumventing the blood–brain barrier, it enables targeted treatment of the central nervous system with remarkable efficiency. The compelling advantages of this approach including low drug dosages, rapid onset of action, and minimal systemic side effects, which make it particularly valuable for pain mechanism research, analgesic drug screening, and investigating neurodegenerative diseases [1, 2]. However, current IT catheter technology still faces limitations, such as complex surgical procedures and catheter blockage, which necessitate the development of improved methods for long-term and reliable application.

Since Yaksh and Rudy [3] pioneered the chronic IT catheter technique in rats in 1976, the field has experienced continuous technological advancements. The early atlanto-occipital approach required inserting the catheter through the foramen magnum and advancing it 8–9 cm into the lumbar region. While this method allowed for precise lumbar drug delivery, the long catheter in the narrow vertebral canal could lead to spinal cord compression, resulting in progressive neurological deficits [2]. Størkson et al. [4] reported that among 40 rats catheterized via the atlanto-occipital route, 2 died and 11 developed neurological dysfunction, whereas no such complications occurred in the lumbar group. Similar concerns about the safety of the atlanto-occipital approach were also shown by Dib [5] and Ineichen et al. [2], who emphasized the advantages of shorter lumbar catheterization routes to minimize neurological injury. By contrast, none of the 32 animals in the direct lumbar group experienced death or neurological dysfunction. Dib [5] also reported complications associated with the atlanto-occipital approach.

To address these serious complications, researchers have explored the direct lumbar puncture approach. Martin et al. [6] were the first to report catheter insertion through the intervertebral space, thereby shortening the IT catheter length, which helps reduce spinal cord compression. Størkson et al. [4] further optimized the puncture procedure by developing the “catheter-through-needle” technique, increasing the success rate to 85%. LoPachin et al. [7] and Milligan et al. [8] made improvements to catheter materials and fixation methods, respectively. In recent years, the 32 G polyurethane (PU) catheter (outer diameter [OD] 0.2–0.25 mm) has gradually become the “gold standard” due to its improved performance [9, 10].

However, existing technologies still face numerous unresolved issues. First, while ultrafine catheters reduce tissue damage, they are highly susceptible to blockage. Jasmin and Ohara [11] reported that, during a 9-month observation period, the functional catheter rate of 32 G PU catheters was 83% (10/12) at 3 months, 58% (7/12) at 6 months, and 50% (6/12) at 9 months. Second, the selection of catheter exit site remains a challenge for researchers. Exit sites on the back or lateral abdomen are easily damaged, requiring single-cage housing, which causes social isolation and adversely affects animal behavior and physiological responses [12, 13]. Hatch et al. [12] described the “isolation syndrome” in rats, and recent studies have confirmed that social isolation can alter pain thresholds and stress responses [13, 14]. Third, existing sealing methods, such as heat sealing and three-way stopcock valves, are complex and costly. Finally, most methods still require 20–30 min of surgical time, which limits their large-scale application [15].

Based on a systematic analysis of existing technologies and a deep understanding of rat spinal cord anatomy, this study proposes an innovative catheter placement strategy. We selected the L5–L6 intervertebral space as the puncture site, which lies below the level of the conus medullaris and contains only the cauda equina, providing a relatively spacious anatomical area [16]. The selection of appropriate sealing methods is critical for maintaining catheter patency while preventing contamination. Although PU catheters have become the mainstream choice [9], this study explored polytetrafluoroethylene (PTFE) catheters for their specific advantages including: (1) low surface energy (18–22 mN/m) minimizing protein adsorption; (2) chemical inertness ensuring drug stability; (3) mechanical properties preventing kinking; and (4) proven biocompatibility in clinical applications [17, 18]. The neck-external design cleverly avoids the problem of catheter damage [11], while the innovative dual-seal closure system ensures both ease of operation and flexibility in usage.

The modified IT catheterization technique established in this study successfully addressed key challenges of traditional methods through systematic innovations. The choice of the L5–L6 approach was based on a precise understanding of rat spinal cord anatomy, whereas the PTFE material provides a balance between ease of operation and long-term efficacy. The neck-external catheter design and dual sealing system demonstrate practical considerations.

The lumbar-sacral enlargement of the rat spinal cord is located at the L3–S1 level [19], and the L5–L6 approach is at the caudal end of the spinal cord, primarily containing the cauda equina. This anatomical feature is the basis for the success of this method. In comparison with the classic method of Yaksh and Rudy [3], we avoided the risk of compression associated with an 8–9 cm long catheter traversing the entire spinal cord. The traditional atlanto-occipital approach may result in neurological complications, including spinal cord compression, nerve root injury, and paralysis [4]. Only 3.3% (1/30) of our rats showed temporary lameness, which completely recovered. Compared to the L4–L5 approach by Martin et al. [6], the L5–L6 position is lower, providing a larger safety margin. Størkson et al. [4] developed the “catheter-through-needle” technique for direct lumbar puncture; however, using the PE-10 polyethylene catheter, we have demonstrated that a thicker catheter, when placed in the appropriate position, is also safe.

The physiological characteristics of the cauda equina enable the use of a 0.5 mm catheter. Histological analyses showed the absence of glial cell activation, with both astrocytes and microglia maintaining their resting morphology, and importantly, neuronal density remained intact with no evidence of neuronal loss in either dorsal or ventral horn regions, which is comparable to the safety profile demonstrated by Sakura et al. [9] using a 32 G catheter. This challenges the long-held belief that “thinner is always better” and suggests that, at specific anatomical sites, the relationship between catheter diameter and damage is not linear. The selection of PTFE was a carefully considered decision. While PU is currently the mainstream material [11], PTFE offers unique advantages. Its low surface energy (approximately 18–22 mN/m) [17, 18] minimizes protein adsorption, which may be key to achieving a 100% high patency rate. By contrast, Jasmin and Ohara [11] reported that, over a 9-month observation period, the functional patency rate of 32 G PU catheters was 83% (10/12) at 3 months, 58% (7/12) at 6 months, and only 50% (6/12) at 9 months. The chemical inertness of PTFE ensures the stability of drugs, which is particularly crucial for protein-based drugs.

The average surgery time of 7.2 min has important practical implications. Several new methods have reported similarly quick operation times, such as a 3–4 min drug delivery time [20], while surgery times for long-term catheter placement vary across different techniques, with lumbar puncture taking 20 min and the atlanto-occipital approach taking 10 min [2]. Our improvements have significantly increased efficiency, which not only reduces the risks associated with anesthesia but more importantly, lowers surgical stress. The body weight gain rate supports this view as rats were only affected in the first week, and quickly recovered thereafter. This means that formal experiments can begin 1 week after catheter placement, greatly reducing the recovery period compared to the 2–3 weeks traditionally recommended. The rapid surgery technique also aids in standardization. We observed a clear learning curve effect (significant time differences between the first and last 10 rats), but even beginners could complete the procedure within 10 min. This reproducibility is crucial for multicenter studies.

The stability of the three pain modalities (mechanical, thermal, and cold) supports the method’s safety. This is consistent with the results of a previous study using fine catheters [21], but our method is simpler to operate. Symmetrical responses from both sides eliminate the possibility of unilateral injury, a factor often overlooked in previous studies. Although this study still required single-cage housing to prevent rats from gnawing on each other’s catheters, the neck-external design has distinct advantages over traditional back or side abdominal exit designs: (1) limited neck mobility, making it difficult for rats to gnaw on the catheter; (2) ease for the researcher to observe the catheter’s condition and administer drugs; and (3) reduced catheter displacement caused by scratching or friction from the rats. Although social isolation has been shown to affect pain thresholds [14, 22], no significant changes in pain sensitivity were observed in this study, possibly due to the rats’ adaptation to single-cage housing conditions. This also suggests that housing conditions may serve as a confounding factor in interpreting pain research results.

The LPS dose-gradient experiment and lidocaine functional verification create a complementary validation framework, confirming the reliability of the catheterization method from different perspectives. The lidocaine experiment confirmed both the accuracy and patency of the catheter placement through its motor block effect, whereas the LPS experiment demonstrated the method’s precision in drug delivery across different doses. The lack of differences in inflammatory markers between the IT-Saline group and the Control group molecularly confirmed the biocompatibility of the catheter 12 weeks after placement, corroborating the histological findings. Furthermore, the distinct dose-dependent gradient effect observed in the three LPS groups strongly supports the dose control precision and repeatability of this catheter technique. This dual verification strategy using both small molecules (lidocaine) and large molecules (LPS) demonstrates the method’s ability to deliver different types of substances in rat models, providing a methodological option for IT drug delivery studies.

The dual closure system design represents a combination of practicality and forward-thinking design. The mechanical compression closure method used in this study performed very well during the 12-week observation period: it only required a hemostatic clamp to operate, with each use consuming 0.5–0.6 mm of the catheter, and the 4 cm reserve theoretically allowed for over 60 uses, fulfilling the requirements for long-term IT drug delivery. This “low-tech” solution is highly accessible in resource-limited settings. With 100% patency and no infections in 30 rats, this method proved highly reliable. Although the infection rate of human IT catheters is relatively high, infection complications are uncommon in rodents [23]. The physical seal formed by mechanical compression blocks contamination at its source—unlike traditional open-ended designs relying on caps or valves—likely being a key factor in the high success rate. For high-frequency drug delivery needs (such as pharmacokinetic studies), we propose the sterile surgical film method as an alternative: inspired by the principle of clinical infusion ports, it allows for long-term repeated drug delivery without shortening the catheter, providing flexible options for different experimental needs.

In recent years, several research groups have attempted to improve the IT catheterization technique. Wu et al. [24] described a simpler and less invasive method for chronic lumbar IT catheterization in mice, which avoids major surgery and causes no neurological deficits. Mazur et al. [15] emphasized rapid catheterization, achieving a success rate of 93%, but still used the traditional PE-10 catheter. Ineichen et al. [2] published detailed protocols in Nature Protocols. Our method combines several innovations: anatomical targeting (L5–L6), material selection (PTFE), external catheter design (neck), and closure system (dual). This systematic improvement has led to significant performance enhancement, especially the 100% long-term patency rate, which is clearly superior to what has been previously reported in the literature.

The catheterization technique established in this study has certain technical limitations. First, the 15 µL dead volume in the catheter affects drug delivery accuracy. With conventional drug doses (15–30 µL), the drug dilution rate is 33–50%. Although this can be compensated by adjusting the drug concentration, for microdosing experiments with less than 10 µL, this limitation will significantly impact experimental precision. This dead volume limitation has specific implications for pharmacological studies. (1) In pharmacokinetic experiments requiring precise drug concentrations, the mixing zone created by the dead volume could alter initial drug distribution patterns. (2) For dose-response studies, especially at the lower end of the curve, the effective delivered dose may deviate significantly from the intended dose. (3) In repeated dosing protocols, residual drug from previous injections could create cumulative effects. Researchers should account for these factors when designing experiments, particularly when comparing results across different catheterization techniques with varying dead volumes. To address this issue, potential technical improvements include: (1) using a catheter with a smaller inner diameter (0.2 mm) to reduce the dead volume to 10 µL; (2) developing a pre-filled catheter system; and (3) designing a dual-lumen catheter for independent control of drug delivery and flushing. Second, the 0.5 mm catheter OD limits its application in certain specialized settings, particularly when the catheter needs to be extended to the thoracic spinal cord. However, based on literature analysis, more than 90% of IT pharmacological studies focus on lumbar administration, so the practical impact of this limitation is relatively small.

To advance the standardized application of this technology, future efforts should primarily focus on: (1) material optimization: enhancing the hydrophilicity of PTFE using plasma treatment and developing PTFE-PU composite catheters that balance biocompatibility with mechanical properties; (2) function expansion: integrating MEMS pressure sensors for patency monitoring and combining with implanted pumps for programmable drug delivery; (3) standardized system: developing a standard operating procedure that includes surgical videos, fault diagnosis flowcharts, and quality control indicators and establishing a multi-center verification mechanism; and (4) cross-species applicability: the neck-external design of this method effectively prevents animals from damaging the catheter, making it suitable for use in other experimental animals such as rabbits and dogs.

The L5–L6 IT catheterization method developed in this study represents a significant technological advancement in the field. Through innovative anatomical localization, material selection, external catheter design, and dual closure system, we achieved rapid (7.2 min), safe (93.3% success rate), and reliable (100% patency rate) long-term IT drug delivery. The key strength of this method lies in its optimal use of rat spinal cord anatomy, where using a relatively thicker PTFE catheter in the cauda equina region is not only safe but may also enhance long-term success rates by minimizing blockages. The neck-external design cleverly addresses the catheter protection issue, while the dual-closure system accommodates various experimental needs. Although the 15 µL dead volume poses a limitation for some precision experiments, it remains an acceptable technical compromise for most IT drug delivery studies. More importantly, this method significantly lowers technical barriers, increases experimental efficiency, and provides a practical tool for neuroscience research, particularly pain and neuropharmacology studies. As this technique is further validated and optimized, it may contribute to more efficient preclinical studies in central nervous system drug development and pain mechanism research.

All data from Western blot (WB) and immunofluorescence assays are available from the corresponding author upon reasonable request. Commercially available experimental reagents are described in detail in the Materials and Methods section. Further inquiries can be directed to zhanghh22@foxmail.com.

The corresponding authors HHZ and WN, as well as the first author WMZ, organized, designed and wrote the article. KZ, YQS, YBD, RR, GHZ and YCM made substantial contributions by refining the study design, performing experiments and data acquisition, conducting data analysis, preparing figures/tables, and conducting literature searches. All authors contributed to editorial changes in the manuscript. All authors have read and approved the manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The experimental protocol involving animals was approved by the Animal Ethics Committee at the Second Hospital of Lanzhou University (Approval No. D2025-459). All animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996) and the ARRIVE 2.0 guidelines.

This work was supported by the National Natural Science Foundation of China (No. 82360435); Lanzhou University Second Hospital Cuiying Youth Fund Project (No. CY2022-QN-A03); Cuiying Science and Technology Innovation Program Project (No. 2022-MS-A10); Lanzhou Youth Science and technology Talents Innovation Project (2023-2-39); Natural Science Foundation of Gansu Province (No. 24JRRA1101).

The authors declare no conflict of interest.

We declare that ChatGPT was used solely as a language editing tool, similar to grammar-checking software or professional editing services, and did not contribute to the intellectual content of this research. All authors have reviewed and approved this declaration.