The dogs were examined under anesthesia at the set postoperative times (6 months and 18 months). In the case of dogs 1–4, the skin was incised along the original incision. The scar outside the dura was stripped off. The dura was incised to separate the cauda equina carefully. The anastomosis was identified by means of the silk thread that was used as a label. For the experimental measurements, the recording electrode was inserted at the distal end of the anastomosis in the right S2 anterior motor root. Meanwhile, the stimulation electrode was inserted at the right L5 posterior sensory root. Both electrodes were connected to the EDJ-V Biosignal system (Shanghai Teaching Equipment Co., Shanghai, China), and unidirectional rectangular pulses of 115 mV were delivered for 1.0 ms. Then, action potential (AP) curves were recorded. For the control measurements, the recording electrode and the stimulation electrode were placed at the left S2 anterior motor root and the left S2 posterior sensory root respectively. The stimulation parameters were the same as those for the experimental measurements.

In the case of dogs 5 and 6, a median incision was made in the lower abdomen. For experimental measurements, the electrodes were inserted into the upper anterior wall of the bladder detrusor muscle, in order to stimulate the right L5 posterior sensory root, the right femoral nerve, and the afferent nerve of the knee-tendon reflex. The electrodes were connected to the Cantata 2000 muscular electrograph (Dantec Medical, Copenhagen, Denmark), and pulses of 3.8 mA were delivered at 1.0 Hz. Electromyograms (EMGs) of the detrusor muscle were recorded. For control measurements, the left S2 posterior sensory root was stimulated with the same parameters as those for the experimental measurements.

In both groups of dogs, paraplegia was induced by transecting the spinal cord acutely at the T10 canal level. Observations were carried out 48 h after paraplegia.

As mentioned before, 48 h after the paraplegia model was induced, the dogs were observed under anesthesia. A median incision was made in the lower abdomen to unveil the bladder cervix and posterior urethra after anesthesia. A 12-F urinary catheter was inserted into the bladder through a small incision on the posterior urethra, after which the incision was sutured. The urinary catheter was then linked to a self-designed urodynamic 3-lead simplified pressure machine [8]. Through the fluid inlet, it was approximately 150 mL of normal saline injected into the bladder. In order to facilitate observing the results, the water column of the pressure test tube was maintained in the baseline of 10-cm. The fluid inlet was turned off. After this, the cauda equina was exposed via the original median approach to look for the previous anastomosis of two anterior motor roots (i.e. the right L5 and S2) in all dog samples. In the cases of Dogs 5 and 6, a right femoral anterior lateral incision was conducted to expose the femoral nerve acted as the afferent nerve of the knee-tendon reflex.

A series of stimulation currents of 1,000 mV at 10 Hz were delivered for 2 s from an EJD-V Biosignal system, in order to stimulate the proximal end of the right L5 anterior motor root anastomosis, right L5 posterior sensory root, left S2 anterior motor root, left L5 posterior sensory root, and right femoral nerve. A medical percussion hammer was used to percuss the right knee tendon and the left knee tendon for 10 s at a frequency of 2 times per second. Increase in bladder pressure (cm H₂O) and duration of urination (T) were recorded. The left S2 anterior motor root served as the normal control, while the left L5 posterior sensory root and left knee tendon were utilized as negative controls.

A nerve specimen was derived from the region 0.5 cm proximal and 1.0 cm distal to the anastomosis, and it was stained with HE in order to count the number of fibers containing medullary nerves with the Leica FW4000 Image Analysis System (Leica, Solms, Germany). The passing rate of myelinic nerve fibers was determined by dividing the number of fibers at the distal end with that at the proximal end.

Data were analyzed using SPSS12.0 and expressed as mean ± SD. The Student’s t-test was used to analyze data. P < 0.05 was considered to indicate statistical significance.