Abstract
Background
Neurogenic bladder (NB) refers to urinary storage and voiding dysfunction resulting from neurological disorders. Animal models of NB are commonly used in preclinical studies, but the distinct characteristics of various modeling techniques are infrequently compared. This study aimed to evaluate and compare the functional and pathological outcomes of different NB rat models.
Methods
Three rat NB models were induced by spinal cord injury (SCI), bladder outlet obstruction (BOO), and diabetes. Blood urea nitrogen (BUN) and serum creatinine (Scr) levels, along with urine flow dynamics, were assessed at weeks 1, 2, 4, and 8. At the 10-week endpoint, animals were euthanized, and bladder weights were recorded for each specimen. Pathological analysis and western blotting were conducted to evaluate bladder muscle fibrosis.
Results
All three rat NB models were successfully established. At week 8, the average maximum/minimum bladder pressures for the SCI, BOO, and diabetic NB rats were 34.0/27.8, 40.4/30.2, and 32.1/28.8 mmH2O, respectively, while bladder capacity and residual volumes were 4.32/4.245, 5.35/5.084, and 4.20/4.048 mL, respectively. The average Scr levels were 52.2, 54.6, and 37.7 mmol/L, and BUN levels were 15.4, 13.8, and 13.9 mmol/L for the three groups. Compared to the control, bladder weights and volumes were significantly increased in the NB groups. Histopathological examination revealed marked thickening and disorganization of the muscle bundles in NB bladders, along with notable inflammatory cell infiltration within the epithelial layer. Immunohistochemical and western blot analyses showed increased fibronectin expression in the NB model bladders.
Conclusions
The three NB rat models effectively replicated clinical and pathological features, including reduced bladder compliance, renal dysfunction, and bladder fibrosis. Among these models, SCI offers the fastest method for inducing NB. Renal function impairment was more pronounced in the SCI- and BOO-induced NB models, with BOO resulting in the most significant pathological changes in the bladder.
Keywords: Neurogenic bladder (NB), animal models, urodynamics, bladder fibrosis
Highlight box.
Key findings
• Spinal cord injury (SCI), bladder outlet obstruction (BOO) and diabetes models each exhibited distinct urodynamic and pathological features. SCI-induced neurogenic bladder (NB) showed rapid deterioration in bladder function within 2 weeks, while BOO-induced NB demonstrated significant bladder fibrosis and stone formation after 4 weeks. The diabetes model displayed a gradual progression of bladder dysfunction over 8 weeks. Renal function impairment, indicated by elevated serum creatinine and blood urea nitrogen levels, was most severe in SCI and BOO models. BOO-induced NB also showed the most pronounced bladder fibrosis and structural changes.
What is known and what is new?
• It is known that NB results from central or peripheral nervous system disorders, leading to bladder dysfunction and renal impairment. Existing models include SCI, BOO, and diabetes-induced NB, but comparative studies on their functional and pathological differences are limited.
• This manuscript provides a detailed comparison of these models, highlighting their unique urodynamic and pathological characteristics, and establishes standardized parameters for modeling.
What is the implication, and what should change now?
• The findings underscore the importance of selecting appropriate NB models based on research objectives. SCI models are suitable for short-term studies, BOO models for advanced NB features, and diabetes models for chronic neuropathy. Future research should focus on refining these models to enhance clinical relevance and explore targeted therapies for NB.
Introduction
Neurogenic bladder (NB) refers to a group of disorders resulting from dysfunction of the central or peripheral nervous system, primarily characterized by impaired urine storage and bladder emptying (1). NB typically presents with significant pathological and functional abnormalities, such as bladder wall thickening and decreased compliance, often leading to elevated bladder pressure and vesicoureteral reflux. Consequently, NB is a major contributor to renal dysfunction, with some patients eventually progressing to end-stage renal disease (2). The prevalence of NB symptoms varies depending on the underlying neurological disorder. For instance, 12–19% of patients with stroke experience NB (3), whereas the incidence can reach up to 90% in those with spinal cord injury (SCI) or multiple sclerosis. Additionally, diabetes-induced peripheral neuropathy can result in NB (4). Given that patients with NB often experience recurrent urinary tract infections, bladder emptying disorders, and even urinary retention, treatment typically involves a combination of pharmacological, surgical, rehabilitative, and nursing interventions (5). Without prompt and effective management, patients’ quality of life and social functioning are significantly impacted, and renal function may deteriorate, threatening both physical and mental health (6).
Pathophysiologically, NB is commonly associated with detrusor-sphincter dyssynergia, where the bladder detrusor and urethral sphincter contract simultaneously during urination (7). This results in elevated bladder pressure and detrusor hyperreflexia, reducing bladder compliance and fostering collagen deposition and bladder fibrosis (8). Bladder fibrosis, characterized by excessive extracellular matrix deposition within the bladder wall, is a pathological alteration commonly seen in chronic inflammation and obstructive lower urinary tract diseases (9). Extensive interstitial or intermuscular collagen deposition diminishes bladder compliance and detrusor function. Current NB treatments, including antimuscarinic drugs, intravesical botulinum toxin injections, and surgery, primarily address symptoms rather than delay the progression of bladder fibrosis (10). Therefore, the successful establishment of an NB model is critical for advancing both fundamental and preclinical research aimed at reversing the pathological effects of bladder fibrosis.
Effective animal models of NB should possess key attributes such as operability, timeliness, and characteristics that closely mimic the human condition. Currently, NB animal models are mainly developed in rats using three primary methods: SCI, bladder outlet obstruction (BOO), and peripheral nerve injury, which is often used to simulate diabetes. SCI disrupts the neural pathways below the spinal cord that control sensory, motor, and autonomic bladder functions, leading to loss of bladder control and an inability to properly coordinate contraction and relaxation during urination, ultimately resulting in NB (11). The absence of neural input for normal bladder control may cause bladder overactivity, manifesting as urgency and frequent urination, which are hallmark symptoms of NB. Furthermore, the neural pathways below the spinal cord also govern sphincter muscle control, and SCI can disrupt sphincter coordination, impairing effective urination (12).
BOO is a condition in no significant urodynamic differences full opening of the bladder outlet, leading to impaired urine flow and, consequently, NB (13). Common causes of outlet obstruction include bladder stones, prostatic hyperplasia, bladder tumors, cystitis, bladder tuberculosis, and pelvic organ prolapse (13-15). BOO causes bladder distension, and over time, the dilation becomes progressively severe, adversely affecting the bladder’s ability to contract, eventually resulting in NB (13). Furthermore, once NB develops, it can exacerbate bladder dysfunction, with bladder distention, trabeculation, and deformity worsening urine retention and further impairing bladder function (16). Additionally, BOO increases the risk of urinary complications (17), with urine retention contributing to recurrent urinary tract infections, bladder stone formation (18), renal function deterioration (19), and other issues, all of which can significantly affect patients’ overall health.
Chronic hyperglycemia in diabetes leads to diabetic neuropathy, often resulting in damage to the autonomic nervous system (20). The autonomic nervous system regulates bladder contraction and relaxation, and its dysfunction can result in abnormal bladder activity, contributing to lower urinary tract symptoms (21). Bladder dysfunction is the most common and persistent urinary issue in diabetic patients (22), with increased residual urine volume being a primary symptom. Recent studies indicate that both urine storage and voiding issues are prevalent in diabetes (23,24). Experimental evidence suggests that the development of diabetic bladder dysfunction is time-dependent (25). This condition progresses from a compensatory state—characterized by bladder enlargement and excessive muscle contractions—to a decompensated state, where muscle contractions become insufficient and urination ability declines, ultimately leading to NB. Additionally, diabetes can cause alterations in bladder sensation and movement, further elevating the risk of NB (26).
Although various approaches exist, there remains a lack of comparative studies and standardized parameters among NB modeling methods, leading to inconsistencies in modeling durations and success rates. This study aimed to address this gap by examining and comparing the functional and pathological outcomes of the aforementioned NB models. We present this article in accordance with the ARRIVE reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-234/rc).
Methods
Animal
A total of 60 female Sprague-Dawley (SD) rats (130–160 g, 6 weeks old), certified specific-pathogen-free (SPF), were obtained from Shanghai Jihui Laboratory Animal Care Co., Ltd. All rats were uniformly fed before the start of surgery to ensure that there were no statistically significant differences among groups at baseline. The rats were housed in an SPF facility under controlled environmental conditions, including temperature (25±2 ℃), humidity (50%±5%), and a 12-hour light/dark cycle. Standard rodent chow and water were provided ad libitum throughout the study. Experiments were performed under project license (No. SHDSYY-2024-1533) granted by the institutional ethics committee of Shanghai Tenth People’s Hospital of Tongji University, in compliance with Chinese national guidelines for the care and use of animals. A protocol was prepared before the study without registration.
Experimental design
The study followed a completely randomized design, treating each animal as an independent experimental unit, consistent with standard interventional study protocols. Sample size determination was based on previous literature (27,28). Sixty female SD rats were randomly assigned to six groups (n=10 per group): SCI sham-operated control, SCI-induced model, BOO sham-operated, BOO-induced model, diabetes model, and diabetes control. No mortality occurred in any experimental group, and the predetermined sample size of n=10 animals per group was maintained for all statistical analyses. No predefined inclusion or exclusion criteria were applied, and no animals or data points were excluded from the final analysis. The number of animals used for each comparison is clearly indicated in the figures and main text. Randomization occurred prior to group assignment; however, investigators were not blinded to group allocation or treatment conditions during the study. This study employed a non-hypothesis-testing approach.
Anesthesia
Prior to surgery, each rat was weighed and anesthetized with 20% urethane (5 mL/kg) via intraperitoneal injection. The surgical procedure was initiated 10 minutes after anesthesia administration.
Construction of SCI-induced NB models secondary to central neuropathy
To construct SCI-induced NB rat models, skin preparation and disinfection were performed along the posterior midline of the 9th thoracic spine (Figure S1A). A skin incision and subcutaneous fascia incision were made, followed by blunt dissection of the erecting spinal muscles to the bilateral sides. The vertebral body was then destroyed to expose the spinal cord, which was clamped with a pressure of 0.2–0.25 N for 15 minutes to induce SCI. After confirming normal breathing and heartbeat, the muscles and skin were sutured using 4-0 silk threads. Sham operations were performed on control rats, in which the spinal cord was exposed but not subjected to clamping.
Construction of NB models secondary to BOO
For the BOO model, a 1-cm incision was made along the midline of the lower abdomen after anesthesia, and the urethra was exposed. The vaginal mucosa was sutured and fixed with 4-0 silk sutures 3 mm below the medial urethra. A small incision was then made on the vaginal wall between the urethra and the sutures. A 3-0 nylon needle with thread was passed through the incision, through the contralateral vaginal wall. The perforated nylon thread was carefully wrapped around the urethra and tied to a 1-mm diameter round steel strip placed in advance. After tying the knot, the steel strip was removed, and the two ends of the nylon thread were drawn out through the vaginal incision and placed in the vagina along with the knot (Figure S1B). The control group underwent only a vaginal wall incision, which was then sutured.
Construction of diabetes-induced NB models secondary to peripheral neuropathy
The SD rats were fasted for 12 hours with free access to water, and streptozotocin (STZ) was administered intraperitoneally based on body weight (35 mg/kg) to induce pancreatic-specific injury. Control rats received an intraperitoneal injection of the same volume of citrate buffer (pH 4.5) without STZ. After 48 hours, random blood glucose levels of the STZ-treated rats exceeded 16.7 mmol/L, indicating successful diabetes induction (Figure S1C). The diabetic group was fed a high-fat diet (45% fat content, Synergy Biology, Nanjing, China), while the control group was maintained on a standard diet.
Urodynamic tests for NB models
Scheduled urodynamic assessments were conducted on all rats at weeks 1, 2, 4, and 8. Three days before the first urodynamic test, a 1-cm incision along the midline of the lower abdomen was made under anesthesia, and a polyethylene catheter with one end sealed by heat was implanted into the bladder dome for both NB and control rats. The catheter was secured with a pouch suture and tunneled subcutaneously to the top of the rat’s head for subsequent urodynamic measurements (Figure S1D-S1F). Three days after catheter implantation, the rats were placed in a metabolic cage while awake. The catheter was connected to a peristaltic pump and data acquisition system (RM-6240, Chengdu Instrument Factory, Chengdu, China). Normal saline at 25 ℃ was continuously infused into the bladder at a rate of 10 mL/h, and urodynamic curves and numerical data were recorded by the acquisition system. The maximum bladder pressure is the highest value during stable fluctuations of intravesical pressure, and the minimum bladder pressure is the lowest value during these stable fluctuations. Bladder capacity is calculated as the time interval from the start of saline infusion until stable pressure is reached, multiplied by the infusion rate of 10 mL/h. After urodynamic measurements, rats spontaneously voided for 45 minutes, followed by manual expression of the bladder. The volume of urine expelled during manual expression was recorded as the residual urine volume.
Assessments of renal function for NB models
Blood urea nitrogen (BUN) and serum creatinine (Scr) levels were measured at weeks 1, 2, 4, and 8. The rats were anesthetized and positioned supine. The chest skin was shaved and disinfected, and the left hand’s index finger and thumb were used to stretch the skin of the rat’s chest and abdomen. A needle was inserted at a 30-degree angle beneath the xiphoid process and advanced to a depth of 2.5–3 cm through the diaphragm. Blood was collected, allowed to clot for 1 hour, and then centrifuged at 3,000 rpm for 20 minutes. The serum was carefully separated and transferred to a new centrifuge tube. BUN and Scr were measured using test kits (Nanjing Institute of Biological Engineering, C011-2-1, C013-2-1), with three duplicates for each sample. Absorbance was determined using an enzyme marker, and BUN and Scr concentrations were calculated based on standard curves.
Prevention of complications
Postoperative care was essential to prevent complications in the NB models of SCI and BOO. On the day following surgery, rats were administered 10 mL of 5% glucose solution orally to provide energy and reduce strain on the digestive and urinary systems. On the first and second days post-surgery, a small amount of solid feed (less than 20 g) was provided, with normal feed resumed on the third day. Sodium lactate Ringer’s solution (20 mL/kg) was administered subcutaneously every 12 hours, and sodium penicillin (200 kU) was injected intraperitoneally every 12 hours from days one to three post-surgery. Penicillin was then continued daily until one week after surgery. Rats in the experimental group were manually assisted with urination and defecation twice daily. Povidone iodine was applied daily to disinfect the wound and the surrounding 5 cm of skin until the wound was fully crusted. The rats’ abdomen and lower limbs were wiped with a 50% ethanol solution to prevent pressure sores, and bitter apple spray was applied near the wounds to prevent self-injury.
Total protein extraction and western blot
Bladder tissue from rats was harvested, and RIPA buffer (Beyotime, Shanghai, China) was added to the homogenization apparatus for complete cell lysis at 4 ℃. The supernatant was then centrifuged at 12,000 rpm, and protein concentration was determined using the bicinchoninic acid (BCA) method (Beyotime), with 40 µg of protein loaded per well. Proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to nitrocellulose membranes (Siema-Aldridge; Merck, Rahway, USA). The membranes were blocked for 1 hour at 25 ℃ with a phosphate-buffered saline (PBS) solution containing 5% skimmed milk powder, followed by overnight incubation at 4 ℃ with the primary antibody (fibronectin, 1:2,000, Proteintech, Wuhan, China). After washing the membranes with phosphate-buffered saline with tween-20 (PBST) three times, they were incubated at 25 ℃ for 1 hour with a secondary antibody (HRP-R, 1:10,000, Proteintech) conjugated to horseradish peroxidase. Protein bands were detected, and their concentrations were quantified using a chemiluminescent imaging system (Tanon5200 System, Tanon, Shanghai, China). Western blot analysis was conducted by a molecular biologist blinded to the experimental group allocations to minimize observer bias.
Pathological examination
The tissues were fixed in cold 4% paraformaldehyde and embedded in paraffin blocks, sectioned into 4-µm slices. Deparaffinization with xylene, gradient alcohol dehydration, antigen retrieval, and sealing were followed by overnight incubation with primary antibodies at 4 ℃. Tissue sections were then incubated with biotinylated goat anti-rabbit immunoglobulin G (IgG) antibody at 25 ℃ for 20 minutes, followed by incubation with streptavidin-horseradish peroxidase for 30 minutes. Finally, diaminobenzidine-H2O2 and hematoxylin were used for tissue staining. Hematoxylin and eosin (H&E) staining was performed after gradient alcohol dehydration, differentiation with 1% hydrochloric acid ethanol, and eosin staining. Masson’s trichrome staining involved deparaffinizing tissue sections, staining nuclei with Weigert’s iron hematoxylin, differentiating with 1% hydrochloric acid alcohol, counterstaining with Masson’s blue solution, staining muscle fibers with a chromic acid and aniline blue mixture, followed by dehydration and mounting with neutral resin. Collagen fibers were visualized in blue, and muscle fibers in red. Histopathological analysis was conducted by a pathologist blinded to experimental group assignments to minimize observer bias.
Statistical analysis
Statistical analysis was performed using SPSS 22.0 software. Measurement data were presented as mean ± standard deviation. For intergroup comparisons, independent samples t-tests were used for pairwise comparisons, while one-way analysis of variance (ANOVA) was employed for comparisons involving three or more groups. Statistical significance was indicated as follows: *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. Results with P≥0.05 were considered not significant (ns).
Results
Successful constructions of three NB models
In this study, 60 rats were used to establish NB models through different methods, including 10 SCI-induced models, 10 BOO-induced models, 10 diabetic models, and 10 controls for each approach. Rats in the three sham-operated groups exhibited clear peak contractions of the detrusor muscle during urination, with maximum bladder pressures recorded at all three time points, followed by a decrease in bladder pressure as urine was expelled. In contrast, rats in the three experimental groups displayed intermittent incontinence. Notably, the maximum bladder capacity was significantly increased, maximum bladder pressure was elevated, and only small amounts of urine were discharged intermittently (Figure S1G-S1I). These results confirm the successful establishment of three distinct NB models. Specifically, all 10 rats in the SCI model group exhibited urinary retention due to SCI, along with bilateral lower limb paralysis. Additionally, they presented with intestinal gas, defecation disorders, reduced food intake, and decreased activity. One rat developed redness and bedsores in the hind limbs. To address these issues, we performed manual bladder and bowel emptying twice daily and applied a 50% ethanol solution to the abdomen and lower limbs of each rat twice daily to improve bedsores. In contrast, no significant complications were observed in the BOO model group. The diabetes model group exhibited polydipsia and polyphagia, but the rats were emaciated.
Three NB models with different urodynamic features
Urodynamic indices, including maximum bladder pressure (Figure 1A and Figure S2A), minimum bladder pressure (Figure 1B and Figure S2B), maximum bladder capacity (Figure 1C and Figure S2C), and residual urine volume (Figure 1D and Figure S2D), were tracked over time. Significant differences were observed in these indices between the NB models and their respective controls. The control groups demonstrated stable urodynamic values over time, with average maximum bladder pressures around 24 mmH2O, minimum bladder pressures at 2.3 mmH2O, bladder capacities at 0.45 mL, and residual urine volumes at 0.03 mL (specific values are provided in Table 1). In contrast, all NB groups exhibited increases in bladder pressures, capacities, and residual volumes (Figure 1A-1D and Figure S2A-S2D). Although bladder pressure decreased during urine excretion, the reduction was minimal, and residual urine volume in the bladder increased significantly, indicating impaired bladder function and compliance. Each NB model displayed distinct urodynamic characteristics. The SCI-induced NB model showed a period of spinal shock lasting approximately one week after surgery, followed by deteriorating urodynamics that stabilized after about 2 weeks (n=10; specific values in Table 1). The BOO-induced model (n=10) reached a stable urodynamic state after 4 weeks. In the diabetic model (n=10), all four urodynamic indices showed a continuous increase over 8 weeks (Figure 1A-1D and Figure S2A-S2D; specific values in Table 1). Interestingly, no significant urodynamic differences were observed between the diabetic model and the control group in the first week; these differences became apparent only in the second week. Notably, at week 8, the BOO-induced model had significantly higher bladder capacity, maximum bladder pressure, and residual urine volume compared to the other two models. While minimum bladder pressures differed among the NB models in the early stages, no significant differences were observed at the 8th week (Figure 1A-1D and Figure S2A-S2D; specific values in Table 1).
Figure 1.
Changes in urine flow dynamics and renal function in NB models. (A) Maximum bladder pressure of different NB models at weeks 1, 2, 4 and 8. (B) Minimum bladder pressure of different NB models at weeks 1, 2, 4 and 8. (C) Maximum bladder capacity of different NB models at weeks 1, 2, 4 and 8. (D) Residual urine volume of different NB models at weeks 1, 2, 4 and 8. (E) Scr levels of different NB models at weeks 1, 2, 4 and 8. (F) BUN levels of different NB models at weeks 1, 2, 4 and 8. Statistical significance was indicated as follows: *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001; ns, results with P≥0.05 were considered not significant. BUN, blood urea nitrogen; BOO, bladder outlet obstruction; NB, neurogenic bladder; SCI, spinal cord injury; Scr, serum creatinine.
Table 1. Mean values of urodynamics and renal function tests at weeks 1, 2, 4, and 8.
| Modeling method | Group | Week | Average of maximum bladder pressure (mmH2O) | Average of minimum bladder pressures (mmH2O) | Average of bladder capacity (mL) | Average of residual urine volumes (mL) | Average of BUN (mmol/L) | Average of Scr (mmol/L) |
|---|---|---|---|---|---|---|---|---|
| SCI | Control | 1 | 24.0 | 2.2 | 0.38 | 0.034 | 7.6 | 24.0 |
| 2 | 24.1 | 2.3 | 0.43 | 0.036 | 8.0 | 24.8 | ||
| 4 | 25.3 | 2.4 | 0.52 | 0.037 | 7.7 | 25.3 | ||
| 8 | 25.6 | 2.5 | 0.60 | 0.035 | 8.2 | 25.1 | ||
| NB | 1 | 25.8 | 17.2 | 2.27 | 1.773 | 7.7 | 24.4 | |
| 2 | 31.7 | 25.0 | 3.75 | 3.997 | 8.5 | 25.0 | ||
| 4 | 33.6 | 26.1 | 4.18 | 4.162 | 11.8 | 32.2 | ||
| 8 | 34.0 | 27.8 | 4.32 | 4.245 | 15.4 | 52.2 | ||
| BOO | Control | 1 | 23.5 | 2.1 | 0.37 | 0.027 | 7.3 | 24.9 |
| 2 | 25.1 | 2.4 | 0.43 | 0.032 | 7.9 | 25.1 | ||
| 4 | 25.7 | 2.6 | 0.52 | 0.031 | 7.6 | 25.3 | ||
| 8 | 26.4 | 2.6 | 0.66 | 0.040 | 8.4 | 24.9 | ||
| NB | 1 | 26.8 | 14.0 | 1.30 | 1.144 | 7.0 | 24.6 | |
| 2 | 31.4 | 19.9 | 2.47 | 2.149 | 8.5 | 24.8 | ||
| 4 | 39.5 | 28.4 | 5.09 | 5.007 | 11.9 | 30.8 | ||
| 8 | 40.4 | 30.2 | 5.35 | 5.084 | 13.8 | 54.6 | ||
| Diabetes | Control | 1 | 21.6 | 2.1 | 0.37 | 0.020 | 7.7 | 24.4 |
| 2 | 22.7 | 2.2 | 0.43 | 0.020 | 8.1 | 25.0 | ||
| 4 | 23.4 | 2.4 | 0.52 | 0.023 | 8.0 | 25.3 | ||
| 8 | 25.3 | 2.6 | 0.60 | 0.031 | 8.3 | 25.2 | ||
| NB | 1 | 22.8 | 2.2 | 0.36 | 0.014 | 8.2 | 24.0 | |
| 2 | 25.2 | 13.3 | 1.29 | 0.787 | 8.9 | 24.8 | ||
| 4 | 29.1 | 21.6 | 3.21 | 2.050 | 11.0 | 27.6 | ||
| 8 | 32.1 | 28.8 | 4.20 | 4.048 | 13.9 | 37.7 |
BUN, blood urea nitrogen; BOO, bladder outlet obstruction; NB, neurogenic bladder; SCI, spinal cord injury; Scr, serum creatinine.
Renal functions of three NB models
Impaired bladder function is frequently accompanied by dysfunction in the upper urinary tract. To assess renal function, blood samples were collected from the rats, and levels of Scr and BUN were measured. All three NB models showed significant elevations in Scr (Figure 1E, Figure S3A, and Table 1) and BUN (Figure 1F, Figure S3B, and Table 1) compared to their respective controls. Although the SCI group (n=10) exhibited the earliest signs of renal dysfunction, no statistically significant difference was observed among the three groups at the 4th week. At the 8th week, the average Scr levels in the SCI (n=10), BOO (n=10), and diabetic (n=10) NB rats were 52.2, 54.6, and 37.7 mmol/L, respectively, while the average Scr levels in the corresponding control groups were 25.1 (n=10), 24.9 (n=10), and 25.2 (n=10) mmol/L. No significant differences were found between the first two NB models, both of which exhibited significantly worse renal function compared to the diabetic model. The average BUN levels of SCI (n=10), BOO (n=10), and diabetic (n=10) NB rats were 14.7, 14.1, and 12.7 mmol/L, respectively, while the average BUN levels of the corresponding control groups were 8.2 (n=10), 8.4 (n=10), and 8.3 (n=10) mmol/L. No significant differences were observed in BUN levels between the groups at any time point. We also performed immunohistochemical (IHC) staining for the renal injury marker KIM-1 to assess the occurrence of renal injury (Figure S3C). The results showed that all rats in the surgical group had essentially the same degree of renal injury.
Pathologic changes of bladder in three NB models
Bladder fibrosis typically develops in the later stages of NB. To capture this, we euthanized the rats at week 10, harvested their bladders, and recorded bladder weight and volume. Consistent with the expected progression, these parameters in all three experimental groups were significantly higher than in their respective controls (Figure 2A-2C). Notably, the BOO-induced model resulted in the largest increases in both bladder weight and volume, compared to the SCI and diabetic models (Figure 2D,2E). H&E staining revealed that the muscle bundles in the bladder tissues of all three NB models were thickened and disordered, with noticeable inflammatory cell infiltration in the epithelial region (Figure S3D). Moreover, bladder stones were even observed in the bladders of rats in the BOO-induced NB group (Figure S3E). Masson’s staining highlighted collagen fibers in blue and smooth muscle cells in red, showing significantly more severe bladder fibrosis in the NB groups than in the control groups, particularly in the BOO-induced model (Figure 3A). Both IHC (Figure 3B) and western blot (Figure 3C) results further demonstrated an increase in fibronectin expression in the bladders of the experimental groups, with the BOO-induced NB rats (n=3) exhibiting the highest fibronectin expression.
Figure 2.
Changes in bladder weight and volume of NB models. (A) Representative bladder images and sagittal H&E staining of the bladder, weight and volume of SCI and control group at the 10-week endpoint. (B) Representative bladder images and sagittal H&E staining of the bladder, weight and volume of BOO and control group at the 10-week endpoint. (C) Representative bladder images and sagittal H&E staining of the bladder, weight and volume of diabetic and control group at the 10-week endpoint. (D) Comparison of bladder weight among the three experimental groups. (E) Comparison of bladder volume among the three experimental groups. Statistical significance was indicated as follows: ****, P<0.0001; ns, results with P≥0.05 were considered not significant. BOO, bladder outlet obstruction; H&E, hematoxylin and eosin; NB, neurogenic bladder; SCI, spinal cord injury.
Figure 3.
Pathologic changes of bladder in NB models. (A) Masson staining of bladder in NB and control groups. (B) IHC staining for fibronectin in bladder tissue of NB and control groups. (C) Western blot analysis of fibronectin expression in bladder tissue of NB and control groups. BOO, bladder outlet obstruction; IHC, immunohistochemical; NB, neurogenic bladder; SCI, spinal cord injury.
Discussion
NB is a disorder of urine storage and bladder emptying caused by dysfunctions in the central or peripheral nervous system (29). The pathogenesis of NB is multifactorial, involving detrusor-urethral sphincter discoordination, leading to elevated bladder pressure, detrusor hyperreflexia, and ultimately, collagen deposition and bladder fibrosis (30). Patients with NB often experience significant pathological and functional impairments, such as bladder wall thickening, reduced compliance, elevated bladder pressure, and vesicoureteral reflux, which may result in renal dysfunction and even progression to end-stage renal disease (31). By comparing three different animal models of NB, this study aims to elucidate their functional and pathological characteristics, which will be instrumental in developing research strategies tailored to NB with diverse etiologies.
In this study, all three NB models demonstrated significant increases in intravesical bladder pressure, though notable differences were observed in the progression and stabilization of these changes. Compared to the BOO model, the SCI model exhibited more rapid deterioration and stabilization of bladder pressure and volume within the first 2 weeks, although the degree of bladder fibrosis was less pronounced. This result may be attributed to the spinal shock period in the SCI model, occurring within one week after surgery, during which the neural pathways are fully disrupted. This disconnection may trigger inflammatory pathways under stress, promoting the infiltration of inflammatory cells (e.g., macrophages, neutrophils) and the release of inflammatory mediators [e.g., tumor necrosis factor alpha (TNF-α), interleukin-1 beta (IL-1β)] (32). These mediators can activate fibroblasts, driving collagen synthesis and deposition, while also exacerbating fibrosis through the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling pathway. After 4 weeks, bladder pressure and volume changes in the SCI model stabilized, and the fibrotic lesions observed could be attributed to the activation of the transforming growth factor beta 1 (TGF-β1) signaling pathway and the sustained inflammatory response caused by chronic urodynamic alterations (33,34). By the 8th week, bladder fibrosis in the SCI model progressively worsened, although the overall pathological changes were less severe compared to the BOO model. This could be due to the SCI model maintaining neurogenic detrusor overactivity primarily through neurogenic factors (e.g., brain-derived neurotrophic factor), while the BOO model is driven largely by mechanical obstruction (35). The BOO model exhibited a significant increase in bladder pressure and capacity within the first 2 weeks post-surgery, though these changes were not as rapid as in the SCI model. This difference may be attributed to the fact that pathological changes in the BOO model are primarily caused by mechanical obstruction, which takes time to accumulate (36). At the 4th week, the BOO model showed a marked increase in bladder capacity, indicating the onset of significant bladder fibrosis. This is likely associated with the activation of the persistent TGF-β1/Smad signaling pathway, the exacerbation of the inflammatory response, and an increase in oxidative stress (37). By the end of the eight-week study, the BOO model displayed the most severe bladder fibrosis. This may be due to the continued chronic inflammation in the bladder, combined with sustained high bladder pressure from mechanical obstruction, which further activates the TGF-β1/Smad signaling pathway, potentially upregulating the expression of tissue inhibitors of metalloproteinases (TIMPs) and inhibiting matrix metalloproteinase (MMP) activity. This process promotes collagen accumulation, exacerbating bladder fibrosis (38). In contrast, the diabetic model showed relatively minor changes in urodynamic parameters and a lower degree of bladder fibrosis. This is likely due to diabetes-induced autonomic neuropathy being a chronic and slowly progressive process, making it difficult to observe significant short-term changes (39). The molecular mechanisms driving bladder fibrosis in this model are likely linked to chronic inflammatory stimulation within the bladder (33).
The extent of renal injury also varied across the models. The SCI and BOO models exhibited more severe renal injury than the diabetic model. The earlier onset of bladder dysfunction in the SCI and BOO models likely led to more substantial urinary retention and vesicoureteral reflux, which increased renal pelvic pressure, contributing to pyelonephritis and renal parenchymal damage. This mechanical pressure directly harms the renal tubules and glomeruli, resulting in decreased renal function (40). Additionally, urinary retention and reflux may trigger local inflammatory responses, releasing inflammatory mediators such as TNF-α and IL-6, which exacerbate renal tissue damage (41,42). These processes also induce oxidative stress, increasing the generation of reactive oxygen species, damaging cell membranes and DNA, and further impairing kidney function (43). In contrast, in the diabetic model, bladder dysfunction appeared later and to a lesser extent. The renal injury in this model may be indirectly related to urinary retention and reflux, as well as microneuropathy (44). Another potential mechanism involves hyperglycemia damaging the kidneys through various pathways, including the accumulation of glycosylation end products, activation of protein kinase C, and activation of the polyol pathway, which can lead to glomerulosclerosis and tubulointerstitial fibrosis (45,46). Additionally, diabetic microangiopathy can reduce glomerular filtration rate, further exacerbating renal impairment (43).
Each NB model presents distinct advantages and limitations. The SCI model is easy to implement with a rapid timeline, but it requires intensive post-operative care to manage bladder function and prevent infections. While the BOO model is more surgically demanding and time-consuming, it offers exceptional accuracy in replicating late-stage NB characteristics. The diabetes model is crucial for studying chronic neuropathic processes, though it requires meticulous experimental control to ensure consistency. These factors guide the selection of models based on specific research goals and experimental context. The translational significance of the three different rat models lies in their ability to provide insights into the difficulty and efficiency of modeling different types of NB. Meanwhile, they can offer preliminary experimental results for future research on the pathophysiology and molecular mechanisms of bladder fibrosis and renal injury in different NB models. However, physiological and anatomical differences between rodents and humans remain a significant challenge for translational applications, highlighting the need for ongoing refinement of modeling techniques to improve both reliability and clinical relevance.
This study has several limitations. First, the 8-week observation period may be insufficient to fully capture the chronic progression of pathological changes, particularly in the diabetic model, which typically requires a longer development time. Second, while fibrosis and urodynamic changes were assessed, the study did not investigate the molecular mechanisms underlying these pathological alterations. Third, potential confounding factors, such as varying activity levels between different model groups and uncontrolled hydration status, were not systematically controlled, which could have impacted bladder and renal function measurements. Fourth, although the discussion explored potential mechanisms leading to bladder fibrosis in the different NB models, our experiments lacked direct comparative results of bladder samples from the models at these time points. These limitations emphasize important areas for future research improvement.
In conclusion, this study offers a comprehensive comparison of three NB models, providing a versatile and effective experimental toolkit for investigating NB with diverse origins. By elucidating the nuanced pathological and mechanistic differences among models, the findings lay a foundation for future research to refine modeling methods, uncover underlying molecular pathways, and translate these insights into transformative clinical therapies.
Conclusions
This study established three NB models—SCI, BOO, and diabetes—which serve as valuable tools for NB research. Each model demonstrates unique characteristics: the SCI model enables rapid construction for short-term studies but requires intensive care; the BOO model closely mimics advanced NB features, such as bladder fibrosis and stone formation; and the diabetes model simulates chronic neuropathy over an extended period. Among these, the BOO-induced NB model exhibited the most severe fibrosis and structural changes, while the SCI and diabetes models showed milder pathological changes but still significant functional impairments. These models provide a solid foundation to guide researchers in selecting the most appropriate method for their studies.
Supplementary
The article’s supplementary files as
Acknowledgments
The authors would like to thank Dr. Xiaotong Qiu, Dr. Yichuan Pang and Quan Zhu for their excellent technical support. The authors thank Bullet Edits Limited for the linguistic editing and proofreading of the manuscript. The image materials for Figure S1A-S1C are sourced from BioRender (www.biorender.com).
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Experiments were performed under project license (No. SHDSYY-2024-1533) granted by the institutional ethics committee of Shanghai Tenth People’s Hospital of Tongji University, in compliance with Chinese national guidelines for the care and use of animals.
Footnotes
Reporting Checklist: The authors have completed the ARRIVE reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-234/rc
Funding: This work was supported by the Experimental Animal Fund of the Shanghai Science and Technology Commission (Nos. 22140903800 and 201409004000).
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-234/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-234/dss
References
- 1.Amarenco G, Sheikh Ismaël S, Chesnel C, et al. Diagnosis and clinical evaluation of neurogenic bladder. Eur J Phys Rehabil Med 2017;53:975-80. 10.23736/S1973-9087.17.04992-9 [DOI] [PubMed] [Google Scholar]
- 2.Stein R, Bogaert G, Dogan HS, et al. EAU/ESPU guidelines on the management of neurogenic bladder in children and adolescent part I diagnostics and conservative treatment. Neurourol Urodyn 2020;39:45-57. 10.1002/nau.24211 [DOI] [PubMed] [Google Scholar]
- 3.Zeng X, Li Q. Effect of Electroacupuncture Treatment on Bladder Function and Urodynamic Features of Neurogenic Bladder in Stroke: A Retrospective Study. Arch Esp Urol 2024;77:1195-201. 10.56434/j.arch.esp.urol.20247710.166 [DOI] [PubMed] [Google Scholar]
- 4.Martens FMJ, Sievert KD. Neurostimulation in neurogenic patients. Curr Opin Urol 2020;30:507-12. 10.1097/MOU.0000000000000773 [DOI] [PubMed] [Google Scholar]
- 5.Panicker JN. Neurogenic Bladder: Epidemiology, Diagnosis, and Management. Semin Neurol 2020;40:569-79. 10.1055/s-0040-1713876 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ginsberg D. The epidemiology and pathophysiology of neurogenic bladder. Am J Manag Care 2013;19:s191-6. [PubMed] [Google Scholar]
- 7.Panicker JN, Fowler CJ, Kessler TM. Lower urinary tract dysfunction in the neurological patient: clinical assessment and management. Lancet Neurol 2015;14:720-32. 10.1016/S1474-4422(15)00070-8 [DOI] [PubMed] [Google Scholar]
- 8.Hackler RH, Hall MK, Zampieri TA. Bladder hypocompliance in the spinal cord injury population. J Urol 1989;141:1390-3. 10.1016/S0022-5347(17)41319-X [DOI] [PubMed] [Google Scholar]
- 9.Lu YT, Tingskov SJ, Djurhuus JC, et al. Can bladder fibrosis in congenital urinary tract obstruction be reversed? J Pediatr Urol 2017;13:574-80. 10.1016/j.jpurol.2017.08.013 [DOI] [PubMed] [Google Scholar]
- 10.DeWitt-Foy ME, Elliott SP. Neurogenic Bladder:: Assessment and Operative Management. Urol Clin North Am 2022;49:519-32. 10.1016/j.ucl.2022.04.010 [DOI] [PubMed] [Google Scholar]
- 11.Hamid R, Averbeck MA, Chiang H, et al. Epidemiology and pathophysiology of neurogenic bladder after spinal cord injury. World J Urol 2018;36:1517-27. 10.1007/s00345-018-2301-z [DOI] [PubMed] [Google Scholar]
- 12.Guo W, Shapiro K, Wang Z, et al. Restoring both continence and micturition after chronic spinal cord injury by pudendal neuromodulation. Exp Neurol 2021;340:113658. 10.1016/j.expneurol.2021.113658 [DOI] [PubMed] [Google Scholar]
- 13.Iguchi N, Hou A, Koul HK, et al. Partial bladder outlet obstruction in mice may cause E-cadherin repression through hypoxia induced pathway. J Urol 2014;192:964-72. 10.1016/j.juro.2014.03.037 [DOI] [PubMed] [Google Scholar]
- 14.Noordhoff TC, Groen J, Scheepe JR, et al. Surgical Management of Anatomic Bladder Outlet Obstruction in Males with Neurogenic Bladder Dysfunction: A Systematic Review. Eur Urol Focus 2019;5:875-86. 10.1016/j.euf.2018.02.009 [DOI] [PubMed] [Google Scholar]
- 15.Utomo E, Groen J, Blok BF. Surgical management of functional bladder outlet obstruction in adults with neurogenic bladder dysfunction. Cochrane Database Syst Rev 2014;2014:CD004927. 10.1002/14651858.CD004927.pub4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ogawa T. Bladder deformities in patients with neurogenic bladder dysfunction. Urol Int 1991;47 Suppl 1:59-62. 10.1159/000282252 [DOI] [PubMed] [Google Scholar]
- 17.Brookman-May S, Burger M, Hoschke B, et al. Association between residual urinary volume and urinary tract infection: prospective trial in 225 male patients. Urologe A 2010;49:1163-8. 10.1007/s00120-010-2364-y [DOI] [PubMed] [Google Scholar]
- 18.Torricelli FC, Mazzucchi E, Danilovic A, et al. Surgical management of bladder stones: literature review. Rev Col Bras Cir 2013;40:227-33. 10.1590/S0100-69912013000300011 [DOI] [PubMed] [Google Scholar]
- 19.Wang Z, Yu Y, Jin L, et al. HucMSC exosomes attenuate partial bladder outlet obstruction-induced renal injury and cell proliferation via the Wnt/β-catenin pathway. Eur J Pharmacol 2023;952:175523. 10.1016/j.ejphar.2023.175523 [DOI] [PubMed] [Google Scholar]
- 20.Mastri AR. Neuropathology of diabetic neurogenic bladder. Ann Intern Med 1980;92:316-8. 10.7326/0003-4819-92-2-316 [DOI] [PubMed] [Google Scholar]
- 21.Podnar S, Vodušek DB. Lower urinary tract dysfunction in patients with peripheral nervous system lesions. Handb Clin Neurol 2015;130:203-24. 10.1016/B978-0-444-63247-0.00012-2 [DOI] [PubMed] [Google Scholar]
- 22.Van Den Eeden SK, Sarma AV, Rutledge BN, et al. Effect of intensive glycemic control and diabetes complications on lower urinary tract symptoms in men with type 1 diabetes: Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study. Diabetes Care 2009;32:664-70. 10.2337/dc07-2375 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Daneshgari F, Liu G, Birder L, et al. Diabetic bladder dysfunction: current translational knowledge. J Urol 2009;182:S18-26. 10.1016/j.juro.2009.08.070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Takewaki F, Hosaka T, Ishitobi M, et al. Characteristics of storage and voiding symptoms in adult patients with type 2 diabetes with lower urinary tract symptoms. J Med Invest 2024;71:237-45. 10.2152/jmi.71.237 [DOI] [PubMed] [Google Scholar]
- 25.Daneshgari F, Liu G, Imrey PB. Time dependent changes in diabetic cystopathy in rats include compensated and decompensated bladder function. J Urol 2006;176:380-6. 10.1016/S0022-5347(06)00582-9 [DOI] [PubMed] [Google Scholar]
- 26.Yuan Z, Tang Z, He C, et al. Diabetic cystopathy: A review. J Diabetes 2015;7:442-7. 10.1111/1753-0407.12272 [DOI] [PubMed] [Google Scholar]
- 27.Li Q, Hong Y, Chen J, et al. Hypoxia-Induced HIF-1α Expression Promotes Neurogenic Bladder Fibrosis via EMT and Pyroptosis. Cells 2022;11:3836. 10.3390/cells11233836 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chen J, Li Q, Hong Y, et al. Inhibition of the NF-κB Signaling Pathway Alleviates Pyroptosis in Bladder Epithelial Cells and Neurogenic Bladder Fibrosis. Int J Mol Sci 2023;24:11160. 10.3390/ijms241311160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Klausner AP, Steers WD. The neurogenic bladder: an update with management strategies for primary care physicians. Med Clin North Am 2011;95:111-20. 10.1016/j.mcna.2010.08.027 [DOI] [PubMed] [Google Scholar]
- 30.Appell RA. Neurogenic bladder. J Urol 1994;152:2086. 10.1016/S0022-5347(17)32315-7 [DOI] [PubMed] [Google Scholar]
- 31.Phé V, Chartier-Kastler E, Panicker JN. Management of neurogenic bladder in patients with multiple sclerosis. Nat Rev Urol 2016;13:275-88. 10.1038/nrurol.2016.53 [DOI] [PubMed] [Google Scholar]
- 32.Yang YD, Yu X, Wang XM, et al. Tanshinone IIA improves functional recovery in spinal cord injury-induced lower urinary tract dysfunction. Neural Regen Res 2017;12:267-75. 10.4103/1673-5374.200810 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Grover S, Srivastava A, Lee R, et al. Role of inflammation in bladder function and interstitial cystitis. Ther Adv Urol 2011;3:19-33. 10.1177/1756287211398255 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ma L, Mu Y, Li X, et al. Expression of transforming growth factor-β1 and autophagy markers in the bladder of rats with neurogenic lower urinary tract injury. Spinal Cord 2023;61:154-9. 10.1038/s41393-022-00866-y [DOI] [PubMed] [Google Scholar]
- 35.Ferreira A, Nascimento D, Cruz CD. Molecular Mechanism Operating in Animal Models of Neurogenic Detrusor Overactivity: A Systematic Review Focusing on Bladder Dysfunction of Neurogenic Origin. Int J Mol Sci 2023;24:3273. 10.3390/ijms24043273 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wang Y, Xiong Z, Gong W, et al. Urethral orifice hyaluronic acid injections: a novel animal model of bladder outlet obstruction. BMC Urol 2015;15:8. 10.1186/s12894-015-0002-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Zhang Y, Gong S, He W, et al. Melatonin serves as a novel treatment in bladder fibrosis through TGF-β1/Smad and EMT. PLoS One 2024;19:e0295104. 10.1371/journal.pone.0295104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Dunton CL, Purves JT, Hughes FM, Jr, et al. BOO induces fibrosis and EMT in urothelial cells which can be recapitulated in vitro through elevated storage and voiding pressure cycles. Int Urol Nephrol 2021;53:2007-18. 10.1007/s11255-021-02942-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Agochukwu-Mmonu N, Pop-Busui R, Wessells H, et al. Autonomic neuropathy and urologic complications in diabetes. Auton Neurosci 2020;229:102736. 10.1016/j.autneu.2020.102736 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Wang ZY, Zhang ZX, Jin LM, et al. Establishment of an Animal Model of Vesicoureteral Reflux Renal Injury through Partial Bladder Outlet Obstruction. Sichuan Da Xue Xue Bao Yi Xue Ban 2021;52:825-31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Wu D, Shi ZE, Xu D, et al. Serum interleukin 6 and acute urinary retention in elderly men with benign prostatic hyperplasia in China: a cross-sectional study. Transl Androl Urol 2021;10:455-65. 10.21037/tau-20-914 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Jang DI, Lee AH, Shin HY, et al. The Role of Tumor Necrosis Factor Alpha (TNF-α) in Autoimmune Disease and Current TNF-α Inhibitors in Therapeutics. Int J Mol Sci 2021;22:2719. 10.3390/ijms22052719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Yu HJ, Lin BR, Lee HS, et al. Sympathetic vesicovascular reflex induced by acute urinary retention evokes proinflammatory and proapoptotic injury in rat liver. Am J Physiol Renal Physiol 2005;288:F1005-14. 10.1152/ajprenal.00223.2004 [DOI] [PubMed] [Google Scholar]
- 44.DiPetrillo K, Coutermarsh B, Gesek FA. Urinary tumor necrosis factor contributes to sodium retention and renal hypertrophy during diabetes. Am J Physiol Renal Physiol 2003;284:F113-21. 10.1152/ajprenal.00026.2002 [DOI] [PubMed] [Google Scholar]
- 45.Noshahr ZS, Salmani H, Khajavi Rad A, et al. Animal Models of Diabetes-Associated Renal Injury. J Diabetes Res 2020;2020:9416419. 10.1155/2020/9416419 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Sowers JR, Epstein M. Diabetes mellitus and associated hypertension, vascular disease, and nephropathy. An update. Hypertension 1995;26:869-79. 10.1161/01.HYP.26.6.869 [DOI] [PubMed] [Google Scholar]