Voiding and the sacral reflex arc: lessons from capsaicin instillation.

Neural control of voiding resides in the pontine micturition centre (PMC), whereas re exes which control bladder emptying may be spinally organised. The efferent and afferent pathways critical for these two bladder functions traverse the spinal cord, and as with all neural processes, sensory input is a critical controlling factor. Interest is now focusing on these afferent mechanisms which determine bladder sensation and in uence the activity of the storage and voiding processes. Studies in man (1, 2) and in the rat (3) have shown that there is a dense plexus of unmyelinated axons in the suburothelium of the bladder, and it is presumed that the majority of these Ž bres are afferent. The full details of the mechanisms whereby bladder sensations (stretch, pain, urgency and the perception of cooling) are generated have yet to be elucidated (4), but de Groat and his colleagues proposed that in the intact cat, afferent impulses are conveyed to higher centres mainly in small myelinated A-delta Ž bres (5). It is now known that there are no direct connections between bladder afferents and the PMC in experimental animals but that on bladder Ž lling there is activation of the periaqueductal grey (PAG) matter (6). Functional imaging in man has also showed activation in the PAG on bladder Ž lling (7). It is thought that activity from there informs higher centres as to the state of bladder fullness and so determines appropriate activation of the PMC. Several weeks after spinal cord damage and disconnection from the PMC, recovery from spinal shock commences and re ex bladder emptying occurs at low volumes and without voluntary control. In the cat, the afferent limb of the re ex arc that causes this activity has been shown to be comprised of unmyelinated C-Ž bres which are normally quiescent in health but become activated following spinal cord injury (5). These afferent Ž bres are sensitive to capsaicin, i.e. they have the vanilloid receptor (VR1) on their surface (8). The presence of the VR1 on a neurone confers a responsiveness such that exposure to a vanilloid such as capsaicin or resiniferatoxin (RTX) results in massive calcium and sodium ion in ow, causing Ž rst excitation then desensitisation followed by cell death (9). Capsaicin is therefore a selective afferent nerve neurotoxin. It was for this reason that in 1991 we Ž rst treated patients with detrusor hyperre exia with strong solutions of intravesical capsaicin (10). For some years, a number of patients with MS and detrusor hyperre exia were successfully treated with repeated instillations of 1 or 2 mmolar capsaicin. Our experience, conŽ rmed by several other groups worldwide, was that not all patients who had what appeared to be a suitable neurological proŽ le, i.e. an incomplete spinal cord lesion with phasic contractions, responded to this treatment (11). In an attempt to examine the mechanism of the response, we took a series of biopsies from patients before and after treatment. Dasgupta established that using a  exible cystoscope, reasonable quality biopsies for both light and electron microscopy containing urothelium, lamina propria and sometimes muscle could be obtained (12). Dasgupta and his colleagues measured the nerve density of PGP 9.5 and S100 staining Ž bres in the lamina propria before and after capsaicin instillations (13). In nine patients who responded, they were able to demonstrate a signiŽ cant reduction in nerve density. This was not the case in the four patients who did not respond. Avelino and Cruz (2000) have since questioned whether the loss of immuno-activity was due to actual axonal degeneration rather than a blockade of axonal transport, the latter being the mechanism they propose for the prolonged response to intravesical vanilloids. The duration of beneŽ t for patients (between 5 and 30 weeks) (14, 15) and the prolonged period of loss of immuno-reactivity in experimental animals (16) certainly suggests more than a transient depletion of neuronal responsiveness.