Translumbosacral Anorectal Magnetic Stimulation Test for Fecal Incontinence

BACKGROUND: Neuropathy may cause fecal incontinence and mixed fecal incontinence/constipation, but its prevalence is unclear, partly due to the lack of comprehensive testing of spino-anorectal innervation. OBJECTIVE: This study aimed to develop and determine the clinical usefulness of a novel test, translumbosacral anorectal magnetic stimulation for fecal incontinence. DESIGN: This observational cohort study was conducted from 2012 to 2018. SETTINGS: This study was performed at a tertiary referral center. PATIENTS: Patients with fecal incontinence, patients with mixed fecal incontinence/constipation, and healthy controls were included. INTERVENTIONS: A translumbosacral anorectal magnetic stimulation test was performed by using an anorectal probe with 4 ring electrodes and magnetic coil, and by stimulating bilateral lumbar and sacral plexuses, uses and recording 8 motor-evoked potentials at anal and rectal sites. MAIN OUTCOME MEASURES: The prevalence of lumbar and/or sacral neuropathy was examined. Secondary outcomes were correlation of neuropathy with anorectal sensorimotor function(s) and morphological changes. RESULTS: We evaluated 220 patients: 144 with fecal incontinence, 76 with mixed fecal incontinence/constipation, and 31 healthy controls. All 8 lumbar and sacral motor-evoked potential latencies were significantly prolonged (p < 0.01) in fecal incontinence and mixed fecal incontinence/constipation groups compared with controls. Neuropathy was patchy and involved 4.0 (3.0) (median (interquartile range)) sites. Lumbar neuropathy was seen in 29% to 65% of the patients in the fecal incontinence group and 22% to 61% of the patients in the mixed fecal incontinence/constipation group, and sacral neuropathy was seen in 24% to 64% and 29% to 61% of these patients. Anal neuropathy was significantly more (p < 0.001) prevalent than rectal neuropathy in both groups. There was no correlation between motor-evoked potential latencies and anal sphincter pressures, rectal sensation, or anal sphincter defects. LIMITATIONS: No comparative analysis with electromyography was performed. CONCLUSION: Lumbar or sacral plexus neuropathy was detected in 40% to 75% of patients with fecal incontinence with a 2-fold greater prevalence at the anal region than the rectum. Lumbosacral neuropathy appears to be an independent mechanism in the pathogenesis of fecal incontinence, unassociated with other sensorimotor dysfunctions. Translumbosacral anorectal magnetic stimulation has a high yield and is a safe and clinically useful neurophysiological test. See Video Abstract at http://links.lww.com/DCR/B728. PRUEBA DE ESTIMULACIÓN MAGNÉTICA TRANSLUMBOSACRAL ANORECTAL PARA LA INCONTINENCIA FECAL ANTECEDENTES: La neuropatía puede causar incontinencia fecal y una combinación de incontinencia fe-cal/estreñimiento, pero su prevalencia no está clara, en parte debido a la falta de pruebas comple-tas de inervación espino-anorrectal. OBJETIVO: Desarrollar y determinar la utilidad clínica de una nueva prueba, estimulación magnética trans-lumbosacral anorrectal para la incontinencia fecal. DISEÑO: Estudio de cohorte observacional del 2012 al 2018. ENTORNO CLINICO: Centro de referencia terciario. PACIENTES: Pacientes con incontinencia fecal, combinación de incontinencia fecal/estreñimiento y controles sanos. INTERVENCIONES: Se realizó una prueba de estimulación magnética translumbosacral anorrectal utilizando una sonda anorrectal con 4 electrodos anulares y bobina magnética, y estimulando los plexos lumbares y sacros bilaterales y registrando ocho potenciales evocados motores las regiones anal y rectal. PRINCIPALES MEDIDAS DE RESULTADO: Se examinó la prevalencia de neuropatía lumbar y/o sacra. Los resultados secundarios fueron la correlación de la neuropatía con las funciones sensitivomotoras anorrectales y cambios morfológi-cos. RESULTADOS: Evaluamos 220 pacientes, 144 con incontinencia fecal, 76 con combinación de incontinencia fe-cal/estreñimiento y 31 sujetos sanos. Las ocho latencias de los potenciales evocadas motoras lum-bares y sacras se prolongaron significativamente (p <0,01) en la incontinencia fecal y el grupo mixto en comparación con los controles. La neuropatía fue irregular y afectaba 4,0 (3,0) (mediana (rango intercuartílico) sitios. Se observó neuropatía lumbar en 29-65% en la incontinencia fecal y 22-61% en el grupo mixto, y neuropatía sacra en 24-64% y 29-61 % de pacientes respectivamen-te. La neuropatía anal fue significativamente más prevalente (p <0,001) que la rectal en ambos grupos. No hubo correlación entre las latencias de los potenciales evocadas motoras y las presio-nes del esfínter anal, la sensación rectal o los defectos del esfínter anal. LIMITACIONES: Sin análisis comparativo con electromiografía. CONCLUSIÓNES: Se detectó neuropatía del plexo lumbar o sacro en el 40-75% de los pacientes con incontinencia fecal con una prevalencia dos veces mayor en la región anal que en el recto. La neuropatía lumbo-sacra parece ser un mecanismo independiente en la patogenia de la incontinencia fecal, no asocia-do con otras disfunciones sensitivomotoras. La estimulación magnética translumbosacral anorrec-tal tiene un alto rendimiento, es una prueba neurofisiológica segura y clínicamente útil. Consulte Video Resumen en http://links.lww.com/DCR/B728.

CONCLUSIÓNES: Se detectó neuropatía del plexo lumbar o sacro en el 40-75% de los pacientes con incontinencia fecal con una prevalencia dos veces mayor en la región anal que en el recto. La neuropatía lumbo-sacra parece ser un mecanismo independiente en la patogenia de la incontinencia fecal, no asocia-do con otras disfunciones sensitivomotoras. La estimulación magnética translumbosacral anorrec-tal tiene un alto rendimiento, es una prueba neurofisiológica segura y clínicamente útil. A norectal disorders, such as fecal incontinence (FI) and mixed FI/constipation are common problems that affect 10% to 15% of the US population. 1,2 Multiple mechanisms contribute to their pathophysiology including sensory and motor dysfunction(s), but whether lumbar and/or sacral neuropathy plays a role remains unclear, 1,3,4 and it has not been systematically assessed. This lack of knowledge stems from technical challenges with current methods of performing clinical neurophysiological tests of the pelvic floor. 5 Neuronal injury causing bowel dysfunction can be assessed by anal electromyography (EMG) or pudendal nerve terminal motor latency testing (PNTML). 4,5 However, needle EMG is painful, requires multiple anal insertions, and is not well tolerated; it provides indirect evidence of neuropathy on the basis of muscle reinnervation. 6 In contrast, PNTML assesses nerve conduction of the distal 2 inches, and was not recommended by the American Gastroenterological Association 7 or the Rome IV committee 4 because it cannot exclude neuropathy and is operator dependent. 4,8 Furthermore, neither test evaluates the entire spino-anorectal neuronal pathways from their origin in the lumbar and sacral plexus to their innervation in the rectum and anal muscles.
A novel approach has been to study the entire spinorectal and spino-anal conduction following magnetic stimulation of the lumbar or sacral plexus. 9-14 When applied topically, over specific sites, in close proximity to the nerves, the magnetic stimulation induces varying pulses of magnetic fields that penetrate up to a depth of 3-to 5-cm, where, on contact with nerves, they induce excitatory postsynaptic potentials. 10,12,15,16 If applied to the lumbar or sacral plexus region, the excitatory postsynaptic potentials evoked in the spinal nerve roots will induce axonal activation of the peripheral nerves that innervate the rectal and anal muscles. If an EMG electrode is located in close proximity to these muscles, the ensuing muscle activation can be recorded as motor-evoked potential. 12,13,16 Thus, magnetic stimulation of the back provides a noninvasive method of examining the lumbar and sacral plexus nerve conduction, and, unlike electrical or needle stimulation, it is well tolerated. 17 Previous small studies using translumbosacral anorectal magnetic stimulation (TAMS) test showed a prolonged motor-evoked potential (MEP) latency in subjects with FI and spinal cord injury, and had a higher diagnostic yield than PNTML. 18,19 However, the prevalence of lumbar and/or sacral neuropathy in anorectal disorders and the clinical usefulness of TAMS for detecting neuropathy are unclear. Also, the relationship between neuropathy and other sensorimotor or morphological dysfunctions of the anorectum is unknown.
Our aims were to: 1) examine the prevalence of lumbar and/or sacral neuropathy in a consecutive series of patients presenting with FI and mixed FI/constipation; 2) determine the clinical usefulness of TAMS test for identifying lumbar and/or sacral neuropathy; and 3) examine the correlation of neuropathy with anorectal motor and sensory function and morphological changes.

METHODS
Consecutive adult patients (≥18 years) referred to the Digestive Health Center between June 2012 and October 2018 with FI (≥2 episodes of liquid or solid stool leakage episodes per month), 1 and mixed FI/constipation (FI = ≥1-2 episodes of liquid or solid stool leakage per month and ≥2 symptoms of Rome III criteria for constipation 20 ) were included in this retrospective observational cohort study. All subjects underwent TAMS test, anorectal manometry, and anal ultrasound using previously described methodology. All patients also filled out a detailed bowel symptoms questionnaire including the frequency of FI episodes, severity of diarrhea, urgency, pain during stooling, straining, incomplete bowel movement, anal itching, abdominal pain, bloating, excess gas and frequency of bowel movement, hard stools, use of digital maneuvers, and enema use. Patients with metal implants (except titanium), brain implantable devices, or previous rectal or anal pouch surgery, or those with inflamed hemorrhoids or active anal fissures were excluded. Data from patients were compared with a normative group of 31 healthy controls. The study was approved by Augusta University Institutional Review Board (1320998).

TAMS Test
With the subject lying in the left lateral position, a 12-cmlong probe (diameter = 8 mm) containing 2 pairs of bipolar steel ring electrodes (Gaeltec Devices Ltd, Dunvegan, Scotland), 1 cm apart, was placed into the anorectum such that the distal pair was located in the anal canal at 1 to 2 cm from the anal verge, and the proximal pair was located in the rectum, at 9-to 10-cm from the anal verge. Also, 3 surface electrodes were placed on the anterior tibialis muscle as control. The magnetic stimulations were performed using the Magstim Rapid2 stimulator (The Magstim Company Limited, Whiteland, Wales, UK). The TAMS test was performed with a 90-mm circular coil (The Magstim Company Limited). Subjects were then turned prone for the TAMS test. The coil was applied on each side at the lumbar and sacral sites, bilaterally, about 4 cm lateral to the midline 18,19 ( Fig. 1). The magnetic stimulations usually began with 50% intensity, and thereafter the intensity was titrated by 5% (up or down) until an adequate and reproducible MEP response, defined as an anal or rectal MEP response with an amplitude of 10 μV and an anterior tibialis MEP response of 20 μV with 3 consecutive trials, was obtained.

Anorectal Manometry and Balloon Expulsion Test
High-resolution anorectal manometry was performed by placing a 12-sensor circumferential solid-state probe (ManoScan AR Catheter, Medtronic, Minneapolis, MN) with a 4-cm-long balloon into the anorectum. Details of the methods used for anorectal manometry have been described previously. [21][22][23]

Anal Rectal Ultrasound
The test was performed with the patient lying in the left lateral decubitus position. A Bruel and Kjaer (Naerum, Denmark) ultrasound transducer (type 2052) with a 360° rotating transducer was placed into the rectum, and anal sphincter morphology, defects, and thickness were assessed. 24

Measurements and Analysis
The latency of the MEP responses obtained at each of the 4 lumbar and 4 sacral sites were measured and compared between groups and controls. Neuropathy was defined as a MEP latency ≥2 SD of the mean value for that site obtained from healthy controls. The prevalence of neuropathy on each side, at each locus, and anal or rectal locations, and in each group of subjects was assessed and compared. We also compared the latencies of MEP and anorectal physiology between patients with FI and mixed FI/constipation and healthy controls. Anal sphincter pressures and rectal sensory thresholds for first sensation, desire to defecate, urge to defecate, and the maximum tolerable volume were recorded. 25 The thickness of the internal anal sphincter (IAS) and external anal sphincter (EAS) and puborectalis, presence of sphincter defects, their location, and degree of defect were assessed, as well as the presence of muscle atrophy or scar. Motor-evoked potential latencies were also compared between patients with or without anal sphincter defects.
The correlation of severity of neuropathy with FI severity was also assessed. We designated the severity of neuropathy based on the number of sites with abnormal MEP latencies; 0 (no), 1 to 2 (mild), 3 to 4 (moderate), and 5 to 8 (severe). The FI severity analysis comprised the frequency of FI episodes.

Statistical Analysis
The quantitative variables (MEP latencies, anorectal physiology parameters) between patients with FI and mixed FI/ constipation and healthy controls were compared using the Mann-Whitney U test. Pearson correlation was used to examine the correlation between age, BMI, duration of symptoms, frequency of FI episodes, anorectal pressures, and rectal sensory thresholds with the number of sites of abnormal MEP latencies. The adjusted R 2 from the logistic regression model was used to evaluate the association between sex, anorectal and back surgery, urine incontinence, types of FI with the number of sites with abnormal MEP latencies.

Demographics and Clinical Characteristics
The demographic features are summarized in Table 1. The prevalence of urinary incontinence, diabetes mellitus, previous anorectal surgery, and back problems were comparable across the 2 groups. The median number of pregnancies in both FI patient groups was 2 and in controls was 1 without group differences ( Table 1). The prevalence of FI was 3-fold higher in women than in men, and the median duration of symptoms varied between 4 and 5 years, but there was no difference between groups. Patients with FI had significantly greater urgency (p = 0.039), higher stool frequency (p < 0.001), and FI episode frequency (p < 0.001) compared with those who had mixed FI/constipation. In contrast, patients with mixed FI/ constipation reported more abdominal pain during stooling (p = 0.003), excessive straining with stooling and feeling of incomplete evacuation (p < 0.001), abdominal pain (p = 0.002) and bloating (p = 0.005), higher frequency of use of digital maneuvers (p < 0.001), and use of enemas (p < 0.001) than patients with FI. The severity of diarrhea, anal itching, excess gas, and passage of hard stools was comparable between the 2 patient groups (p > 0.107; Table 1).

Assessment of Lumbar Plexus Neuropathy
Typical anal MEP response from a healthy subject and patients with FI and mixed FI/constipation is shown in Figure 2. Bilateral translumboanal and translumborectal MEP latencies in both patient groups were significantly prolonged (p < 0.001) compared with controls (Table 2). Also, 55.3% to 65.3% had left-or right-sided lumboanal neuropathy, and its prevalence was significantly higher (p < 0.001) than lumborectal neuropathy in both groups (Fig. 3). In subjects with FI, neuropathy was patchy and, on average, involved 2.0 (2.0)/4 loci. Also, 31 of 144 (21.5%) had neuropathy affecting all 4 loci. In patients with mixed FI/constipation, on average, 1.5 (1.8)/4 loci were involved and 9 of 76 (11.8%) had neuropathy affecting all 4 loci. There was no difference between groups (p > 0.076).

Assessment of Sacral Plexus Neuropathy
Bilateral transsacral-anal and transsacral-rectal MEP latencies were significantly prolonged in patients with FI and mixed FI/constipation compared with controls (p < 0.05) and, on average, involved 2.0 (1.0)/4 loci in the FI group as well as 2.0 (2.0)/4 loci in the mixed FI/constipation group. Also, the prevalence of transsacral-rectal neuropathy was significantly lower (p < 0.01) than transsacral-anal neuropathy in both groups (Fig. 3).

Anorectal Manometry
The resting anal and maximum and sustained squeeze sphincter pressures were significantly lower (p < 0.001) in patients with FI and mixed FI/constipation than in healthy controls, but there were no differences between patient groups (Table 3). When bearing down, the anal residual pressure was significantly higher (p < 0.01), and the defecation index was significantly lower (p<0.001) in both patient groups than in healthy controls. The rectal sensory thresholds for first sensation and constant sensation were similar between both patient groups compared with healthy controls (p > 0.05). The threshold for desire to defecate, urge to defecate, and maximum tolerable volume were significantly lower in the mixed FI/constipation group than in healthy controls (Table 3). Only the maximum tolerable volume was significantly lower in the FI group than in healthy controls. The mean balloon expulsion time (BET) was significantly prolonged in patients with FI (p < 0.05) and patients with mixed FI/constipation (p < 0.001) compared with healthy controls. Significantly more patients in the mixed FI/constipation group had abnormal BET (p = 0.036) and dyssynergic defecation (p = 0.019) than in the FI group.
number of loci with abnormal MEP (p > 0.05; Table 4). The maximal squeeze pressures in the mixed FI/constipation group were slightly negatively associated with the number of loci of abnormal MEP (r > -0.269; p < 0.0265; Table 4).

Anal Rectal Ultrasound
The distribution of external or internal anal sphincter defect is summarized in Tables 1 and 5. The prevalence of EAS or IAS sphincter defects or both sphincter defects was similar between the FI and mixed FI/constipation groups. In addition, bilateral lumbar and sacral anorectal MEP latencies were not significantly different in patients with or without anal sphincter defects (p > 0.05), except bilateral lumbar-anal and right sacralanal latencies in mixed FI/constipation groups (Table 5).

Factors Associated With the Number of Loci and Abnormal MEP Latencies
The demographic and clinical characteristics including age, sex, symptom duration, number of births and vaginal deliveries, fecal incontinence types (passive, urgency, and both), and the presence of diabetes mellitus, urinary incontinence, back injury/back surgery, and frequency of FI episodes were not correlated with the number of the loci of abnormal MEP latencies in both groups (p > 0.05), except for female sex, which was more likely to have a greater number of sites with  abnormal MEP latencies in both groups (r > 0.162; p < 0.046). Also, higher BMI was more likely to have more sites with abnormal MEP latencies in the FI group, but not in the mixed FI/ constipation group. There was a significant correlation between anorectal and pelvic surgery and sites of abnormal MEP latencies in the FI group (r > 0.193; p < 0.021), but not in the mixed FI/constipation group (p > 0.660; Table 4).

Neuropathy Severity and Correlation With FI Severity
Based on the number of loci of abnormal MEP latencies, the severity of neuropathy in FI and mixed FI/constipations groups was categorized as normal (7.6% versus 7.9%), mild (23.6% versus 27.6%), moderate (37.5% versus 34.2%), and severe (31.3% versus 30.3%). Overall, there was no difference (p = 0.92). In addition, there was no association between the severity of neuropathy and frequency of FI episodes (r = 0.026; p = 0.708) in patients with FI or mixed FI/constipation.

Technical Limitation
Although TAMS data were obtained from all patients in this study, in 3 patients, data were obtained from only 7 sites, in 7 patients from only 6 sites, in 3 patients from only 5 sites, and in 1 patient from only 4 sites. Reasons for these technical limitations include obesity (n = 6), back injury/ surgery (n = 2), and equipment failure (n = 6).

DISCUSSION
We found that a significant number of patients with FI and mixed FI/constipation demonstrated either lumbar plexus or sacral plexus neuropathy affecting the anorectum. Also, the prevalence of neuropathy in the anal region was 2-fold higher than in the rectal region, and in both patient groups. The prevalence of either lumbar and/or sacral plexus neuropathy, however, was similar in both groups of patients. Also, age, duration of illness, number of births, prevalence of back injury and surgery, diabetes mellitus, frequency of FI episodes, and most anorectal physiological parameters were not associated with the number of loci with neuropathy. However, patients with higher BMI, women, and patients with previous anorectal and pelvic surgery had weak but significant correlation for the involvement of more sites with lumbosacral neuropathy. The coexistence of FI and constipation is well known in the pediatric and geriatric population. 26 In a series of 4027 patients, over 40% of adults had mixed FI/constipation, 27 but their pathophysiology is poorly understood.  Hence, we separated the 2 groups to better understand the underlying pathophysiology and to facilitate treatment. We found a higher prevalence of neuropathy in the mixed FI/constipation group, and this was comparable to the FI group, suggesting that neuropathy is an important factor in the pathogenesis of their symptoms. The higher prevalence of dyssynergic defecation and prolonged BET may contribute to FI and/or seepage alongside neuropathy in this mixed group. 28 Although the precise anal or rectal pelvic nerves that are stimulated with the TAMS test are unclear, the measurements of anorectal MEP latencies usually represent a summation of conduction times of motor nerves from the pyramidal pathway, along with the activation of α-motor neurons. 29 We believe that the lumbar site of stimulation most likely activates the L2 and L3 nerves including the superior hypogastric plexus, the lumbosacral trunk, lumbar splanchnic nerve, and genitofemoral nerves. Likewise, A scale of 0 (never), 1 (less than 2/week), 2 (2/week), 3 (1/day), 4 (more than 3/day). the sacral site of stimulation, likely activates the S2, S3, and S4 nerves, including the pudendal nerve, the inferior hypogastric plexus, and the pelvic and sacral splanchnic nerves. The number of sites with abnormal MEP latencies was not associated with anorectal sensorimotor dysfunction in either FI or mixed FI/constipation. This not only confirms previous studies in FI, 18,30 but extends those observations to the cohort of patients with mixed FI/constipation. Previous studies that have assessed the correlation of anorectal neuropathy using PNTML with anal sphincter function have found mixed results. Súilleabháin et al 31 showed that the maximum squeeze pressure was not significantly correlated with PNTML latency, whereas Saraidaridis et al 32 showed lower median maximum squeeze pressure with prolonged PNTML but not normal PNTML. Loganathan et al 33 suggested decreased resting and squeeze pressures with unilateral or bilateral prolonged PNTMLs. These observations may be due to the use of different inclusion criteria and/or issues with PNTMLs, including operator variability and poor sensitivity and specificity. 34 Hence, it is important to use a more objective and comprehensive diagnostic test that assesses the entire neural pathway from their origin in the spinal cord to its innervation in the rectum and anal canal. Also, TAMS test has been shown to be superior to PNTML regarding sensitivity and diagnostic yield. 18,19 We found that the anal and rectal neuropathy were not significantly associated with anal sphincter defects in both groups, except weak association of bilateral lumboanal and right sacral anal latencies in the mixed FI/constipation group with anal sphincter defects. Thus, the results of this large group of patients with FI are consistent with our smaller previous study that showed no significant correlation between MEP latencies and anal sphincter morphology (EAS or IAS defects and thickness). 35 Consequently, anal and rectal neuropathies appear to be independent factors contributing to the pathophysiology of anorectal dysfunction and are not always correlated with sensorimotor dysfunction(s).
The limitations of this study include the recruitment of patients from a tertiary center that limits generalizability and the lack of comparative assessment with either EMG or PNTML to independently verify lumbosacral neuropathy. However, previously, we showed that the yield of neuropathy was significantly lower with PNTML compared with TAMS, 18,19 and whenever neuropathy was detected with PNTML, it was also identified by TAMS. 18,19 Hence, comparative assessment is less likely to provide additional information. Also, there was an age difference between patient groups and controls, but age did not appear to influence MEP latencies. 36 The weak but significant correlation of factors such as women, obesity, and previous anorectal as well as pelvic surgery with the number of sites showing lumbosacral neuropathy suggest that these factors may be associated with pelvic floor neuropathy.

CONCLUSION
In conclusion, TAMS appears to be a useful, safe, comprehensive, noninvasive, neurophysiological test for detecting lumbar or sacral plexus neuropathy affecting anal or rectal region in patients with FI. Neuropathy was detected in a high proportion of patients who had FI and mixed FI/constipation, reaffirming its importance as an underlying mechanism and an independent risk factor. The identification of neuropathy could pave the way for neuromodulation therapies such as sacral nerve stimulation 37 and translumbosacral neuromodulation therapy. 38 Given its high diagnostic yield, TAMS should be a preferred test for the neurophysiological assessment of anorectal disorders.