Objective To study the microstructural change of detrusor muscle and neuromuscular junction (NMJ) after bladder functional reconstruction for atonic bladder caused by medullary cone injury and to discuss the feasibility of bladder functional reconstruction for improving the detrusor muscle degeneration. Methods A total of 104 adult female Sprague-Dawley rats (weighing, 200-250 g) were randomized divided into 3 groups: normal group (n=8), control group (n=48), and experimental group (n=48). No treatment was given in normal group; the medullary cone injury was established by sharp transection of spinal cord at L4, 5 levels in control group; and the anastomosis of bilateral L5 ventral root (VR)-S2 VR and L5 dorsal root (DR)-S2 DR was performed for bladder functional reconstruction after modeling of medullary cone injury in experimental group. After operation, the survival condition of rats was observed. At 3 days and 3 consecutive days before 1, 2, 3, 4, 5, and 6 months after operation, the residual urine volume was measured; at 1, 2, 3, 4, 5, and 6 months after operation, the detrusor muscle was harvested to measure the muscle fiber cross-sectional area by HE staining, to calculate the percentage of connective tissue by Masson trichrome staining, and to observe the ultrastructure of the detrusor muscle and the NMJ by transmission electron microscope (TEM). Results Eleven rats were supplemented because of death after operation. In control group, a significant increase of the residual urine volume was observed with the extension of time (P lt; 0.05); in experimental group, an increase was observed at the first 3 months after operation, and then gradually decreased, showing significant differences between the other time point (P lt; 0.05) except between at 3 days and at 5 months after operation (P gt; 0.05); there was significant difference between control and experimental groups at other each time point (P lt; 0.05) except at 3 days, 1 month, and 2 months (P gt; 0.05). HE staining and Masson trichrome staining indicated that the muscle fibers arranged in disorder with gradually aggravated atrophy and gradually increased connective tissue in control group, while the shape of the detrusor muscle recovered with no increased connective tissue at 4, 5, and 6 months after operation in experimental group; there was significant difference in cross-sectional area of detrusor muscle and percentage of connective tissue between normal group and experimental group, and between normal group and control group at each time point (P lt; 0.05). In control group, the cross-sectional area of detrusor muscle decreased and the percentage of connective tissue increased with the extension of time (P lt; 0.05). In experimental group, the cross-sectional area of detrusor muscle decreased at the first 3 months and then increased, and the percentage of connective tissue increased slowly with the extension of time. There was no significant difference of cross-sectional area of detrusor muscle at the first 3 months between control and experimental groups (P gt; 0.05), but the values in experimental group were significantly higher than those in control group at 4, 5, and 6 months after operation (P lt; 0.05). There were significant differences of the percentage of connective tissue between control and experimental groups at each time point (P lt; 0.05). In control group, the amount of synaptic vesicles decreased in the NMJ with time passing; vacuole like structure was observed in NMJ at 3 months; there was almost no nerve ending at 6 months. In experimental group, the amount of synaptic vesicles decreased at 1 and 3 months after operation, but obviously increased at 6 months. Conclusion The reconstruction of bladder function with L5 nerve roots above the paraplegic plane can effectively inhibit the degeneration of detrusor muscle and improve its microstructural changes after medullary cone injury.