Are White Matter Tract Integrities Different in Multiple Sclerosis Women With Voiding Dysfunction?

Abstract

Objectives:

Two white matter tracts (WMTs) are proposed to be involved in bladder function: anterior thalamic radiation and superior longitudinal fasciculus. Multiple sclerosis (MS) patients with voiding dysfunction (VD) may have distinct changes in these 2 WMTs. This study aims to compare the fractional anisotropy (FA) and mean diffusivity (MD) from diffusion tensor imaging of MS females with and without VD versus healthy controls (HCs).

Methods:

Prospective observational cohorts of 28 female MS patients and 11 HCs were recruited. Multiple sclerosis patients were divided into 2 groups: voiders (patients without VD, n = 14) and VD (patients with VD, n = 14). Diffusion tensor imaging of each subject was obtained, from which FA and MD maps were generated. The mean FA and MD of each WMT on both sides were analyzed using one-way analysis of variance and pairwise comparison with adjusted P values.

Results:

Overall MS patients had significantly lower mean FA (loss of coherence) and significantly higher mean MD (increased free diffusion) than HCs in both WMTs, indicating more damage. Furthermore, VD showed a trend of loss of integrity in both WMTs when compared with voiders with lower FA and higher MD.

Conclusions:

There is damage reflected by lower FA and higher MD values in the proposed WMTs involved in bladder function in MS women. Voiding dysfunction in this patient population can be attributed to these damages considering women with VD demonstrated a trend of deterioration in these WMTs compared with women without VD. Future studies with larger sample sizes should be done to further confirm this correlation.

Voiding dysfunction (VD) is a morbid and costly urological condition characterized by hesitancy, intermittency, or absence of flow of urine,¹ leading to incomplete emptying of the bladder, urinary retention, urinary tract infections, stones, or permanent renal failure.² Normal voluntary voiding is mediated by the suprapontine centers in the brain. The pontine micturition center triggers (switches on) the voiding reflex, in which the urethral sphincter/pelvic floor relaxes and the detrusor muscle contracts.³ Voiding dysfunction often occurs if the detrusor muscle cannot initiate or maintain effective contractions (ie, detrusor underactivity), the urethra fails to relax and lower urethral resistance, or there is failure in synchronizing these actions.^2,4 Voiding dysfunction is one of the most common symptoms in patients with neurologic disorders, such as multiple sclerosis (MS)^5,6—a chronic autoimmune, inflammatory disease where axons in the central nervous system are demyelinated, thus impairing the nerve signaling.⁷

With the evolution of neuroimaging tools and resources such as functional magnetic resonance imaging (fMRI), we are gaining insight into how the brain controls bladder function in healthy individuals. Our understanding of the involvement of brain centers and their connections in relation to VD, especially in neurogenic patients, is still limited.⁸ This study aims to assess the integrity of 2 white matter tracts (WMTs) proposed to be involved in proper lower urinary tract function—the anterior thalamic radiation (ATR) and superior longitudinal fasciculus (SLF)^5,9 and their correlation with VD in MS women. These tracts were chosen to be evaluated as they were found to connect several regions that are activated and deactivated by bladder filling.^5,9,10 In addition, the arcuate fasciculus (AF), a commonly damaged tract in MS patients that has been found to be involved in the language processing pathway,^11–13 but not in the bladder cycle, is also evaluated to serve as a control WMT to confirm that the deterioration of the ATR and SLF are related to voiding in MS patients and not to the global advancement of the disease. We hypothesize that there is more damage in the ATR and SLF tracts of MS women with VD when compared with MS women without VD or healthy controls (HCs), and in the AF of MS women when compared with HCs but not between the 2 MS groups. To test our hypothesis, we compared the fractional anisotropy (FA) and mean diffusivity (MD) of the WMTs on diffusion tensor imaging (DTI) brain scans in MS women with and without VD, and HCs. Because the roles of WMTs in humans with VD, especially neuropathic patients, have not been investigated previously, the result of this study is seminal in identifying diagnostic, prognostic, and potentially therapeutic measures for VD in MS patients.

MATERIALS AND METHODS

Subjects

Twenty-eight female MS patients and 11 female HCs were recruited for this study. Multiple sclerosis patients were divided into 2 groups based on their cystometric studies during their urodynamic study (UDS) visit performed at our tertiary care center: patients without VD (n = 14) who void spontaneously and have a postvoid residual (PVR) urine volume of less than 40% maximum cystometric capacity (MCC), and patients with VD (n = 14) who perform self-catheterization or have a PVR volume of 40% or more MCC. Adult female patients with clinically stable MS for 6 months or more with symptomatic neurogenic lower urinary tract disorders (NLUTDs) were referred to fellowship-trained neurourologists at our tertiary care center. All patients completed a detailed history, physical examination, and validated questionnaires that included the Urogenital Distress Inventory (UDI)-6, Incontinence Impact Questionnaire 7, MRI Safety Screening Questionnaire, and Hamilton Anxiety Rating Scale.

Patients’ demographics with P values from unpaired t tests and χ² tests (α = 0.05) are detailed in Table 1. There were significant differences in age, questionnaires, voiding patterns, overactive bladder medication, PVR, and %PVR/MCC between MS patients and HCs, indicating the expected presence of the pathology in MS patients. Significant difference was found only in voiding patterns, PVR and %PVR/MCC between voiders and VD groups, which was the basis for our group division.

TABLE 1.

Patients’ Demographics

	Voiders	VD	P (Voiders vs VD)	Overall MS	Healthy Control	P (MS vs HC)
No. patients	14	14		28	11
Age, mean (SD)	49.6 (9.8)	52.9 (15.6)	0.521	51.3 (12.9)	33.0 (6.1)	<0.001*
BMI, mean (SD)	29.9 (7.0)	26.5 (4.6)	0.147	28.2 (6.1)	25.5 (6.4)	0.249
Duration of MS, mean (SD), y	15.1 (10.8)	16.5 (13.7)	0.773	15.8 (12.1)	NA	NA
UDI-6, mean (SD)	10.9 (7.2)	11.6 (4.4)	0.778	11.3 (5.9)	0.1 (0.3)	<0.001*
UDI-6, Q5 (voiding)	2.3 (1.6)	3.4 (1.2)	0.052	2.8 (1.5)	0 (0)	<0.001*
IIQ-7, mean (SD)	7.9 (7.7)	9.2 (6.1)	0.629	8.6 (6.8)	0 (0.3)	<0.001*
Voiding patterns, n (%)
Strictly voiding spontaneously	14 (100)	4 (28.6)	<0.001*	18 (64.3)	11 (100)	0.022*
Strictly on self-catheterization	0 (0)	10 (71.4)	<0.001*	10 (35.7)	0 (0)	0.022*
Voiding spontaneously and on self-catheterization	0 (0)	2 (14.3)	0.142	2 (7.1)	0 (0)	0.363
Previous hysterectomy, n (%)	3 (21.4)	2 (14.3)	0.622	5 (27.8)	0 (0)	0.133
Overactive bladder medication use at baseline, n (%)	12 (85.7)	12 (85.7)	1.000	24 (85.7)	0 (0)	<0.001*
Urodynamic data, n (%)
Mean MCC, mL	412.6 (163.7)	363.6 (164.5)	0.437	388.1 (163.0)	369.9 (66.1)	0.623
PVR, mL	49.5 (48.3)	206.6 (109.3)	<0.001*	128.0 (115.2)	16.0 (24.9)	<0.001*
% PVR/MCC	13.1 (11.5)	57 (22.9)	<0.001*	35 (28.6)	3.9 (5.9)	<0.001*
Detrusor sphincter dyssynergia	1 (7.1)	3 (21.4)	0.280	4 (14.3)	0 (0)	0.186
Baseline MRI findings, n (%)					NA	NA
Presence of general cortical atrophy	3 (21.4)	3 (21.4)	1.000	6 (21.4)
Presence of enhancing lesions	1 (7.1)	3 (21.4)	0.280	4 (14.3)
Location of lesions
Cerebrum	14 (100)	14 (100)	1.000	28 (100)
Cerebellum	2 (14.2)	6 (42.9)	0.094	8 (28.6)
Brainstem	7 (50)	4 (28.6)	0.246	11 (39.2)
Spinal cord	8 (57.1)	7 (50)	0.705	15 (53.6)

Unpaired t test and χ² tests (α = 0.05) were performed on continuous and binary variables, respectively, to compare voiders and VD groups, and all MS and HC groups.

^*Significant difference between groups (P < 0.05).

Abbreviation: IIQ-7, Incontinence Impact Questionnaire 7.

Diffusion Tensor Imaging Paradigm

Diffusion tensor imaging is a powerful fMRI-based neuroimaging technique that allows WMTs to be characterized by their location, direction, and integrity. Because of the architecture of axons in parallel bundles and their myelin sheath, water molecules naturally diffuse along the axon’s main direction.¹⁴ Diffusion tensor imaging allows the diffusion tensor to be calculated for each axon fiber tract, which can then be used to compute 2 metrics, FA and MD, of each WMT. Deficits in WMTs are indicated by lower FA and higher MD values. Fractional anisotropy describes the main direction of diffusion, and higher FA values indicate higher signal transmission of the fiber tracts along their main axis.^14,15 Mean diffusivity, on the other hand, indicates water diffusion in the brain tissues’ extracellular space, independently from the tract direction.^15,16 Higher MD values, therefore, indicate an increase in free diffusion or dispersion of signal. Assessing these 2 measures could provide further insights into the correlation between the integrity of the WMTs and VD in MS patients.

Our previously established concomitant UDS/fMRI platform was used in this study to obtain images during filling and the voiding phases of the micturition cycle.¹⁷ Diffusion tensor imaging of each subject was obtained before the UDS/fMRI platform.

Data Acquisition and Statistical Analysis

A schematic of the process is detailed in Supplementary Appendix 1 http://links.lww.com/FPMRS/A117. Briefly, dcm2nii¹⁸ was used to obtain the B values and gradient directions for each DTI. FMRIB Software Library (FSL)¹⁹ was used to strip the skull, create a brain mask, and fit the data into diffusion tensors to generate FA and MD maps of the brain. The FA, MD, and color FA maps (generated using Diffusion Toolkit²⁰) were then imported into TrackVis.²⁰ To identify the tracts, a spherical region of interest was used on the color FA map to estimate the spatial location for each tract according to literature. The tract was defined as all fibers running through these regions of interest (Fig. 1). After the tracts were located, their FA and MD were obtained through TrackVis. One-way analysis of variance (ANOVA) (α = 0.05) was performed to detect significant difference in FA and MD of the WMTs on both sides between all 3 groups (voiders, VD, and HCs). Pairwise comparison t test (α = 0.05) with adjusted P values using the Benjamini-Hochberg method was performed subsequently to compare between the groups and between all MS patients and HCs.

FIGURE 1.

Anatomy of the WMTs of interest (demonstrated on subject 18). A, axial and B, Sagittal ATR views. Axial (C) and sagittal (D) SLF views. Colors indicate direction of fiber tracts (red = transverse, blue = craniocaudal, green = anterior posterior).

RESULTS

Diffusion Tensor Imaging Analysis Between VD, Voiders, and HCs

Fractional Anisotropy

Two-sample t test showed that MS patients overall demonstrated statistically significant lower FA values than HCs in all 3 WMTs on both sides (P < 0.05). One-way ANOVA showed statistically significant difference in all WMTs among all 3 groups. Specifically, pairwise t test indicated that MS women with VD showed significantly lower FA values than HCs in the ATR and SLF on both sides. When compared with voiders, VD group demonstrated a trend of lower FA in all tracts (P > 0.05) (Fig. 2, Table in Supplemental Digital Content [SDC] 1 http://links.lww.com/FPMRS/A117). Lastly, no significant difference was found between voiders and VD in the left and right AF (control tract) (Fig. S3 http://links.lww.com/FPMRS/A119).

FIGURE 2.

Fractional anisotropy value comparison of the left ATR, right ATR, left SLF, and right SLF among voiders, VD, all MS patients, and HC groups. Pairwise comparisons with an adjusted P value using the Benjamini-Hochberg method (α = 0.05) were performed on voiders, VD, and HC groups, and 2-sample t tests (α = 0.05) were performed on all MS and HC groups. Bars indicate standard deviation, and asterisks indicate significant difference between groups (P < 0.05). Fractional anisotropy values of all tracts in VD group were lower than those in voiders group.

Mean Diffusivity

Two-sample t test on MD also showed that overall MS patients demonstrated significantly higher MD values compared with HCs in all 3 WMTs on both sides (P < 0.05). However, one-way ANOVA only indicated statistically significant difference among 3 groups in the left ATR. Pairwise t test showed that MS women with VD have higher MD value in the left ATR when compared with HCs with significant difference (P = 0.049) and with voiders with difference approaching significance (P = 0.054). In addition, VD showed a trend of higher MD in all tracts compared with voiders (P > 0.05) (Fig. 3, Table S2 http://links.lww.com/FPMRS/A118). Lastly, no significant difference was found between voiders and VD in the left and right AF (Fig. S3 http://links.lww.com/FPMRS/A119).

FIGURE 3.

Mean diffusivity value comparison of the left ATR, right ATR, left SLF, and right SLF among voiders, VD, all MS patients, and HC groups. Pairwise comparisons with an adjusted P value using the Benjamini-Hochberg method (α = 0.05) were performed on voiders, VD, and HC groups, and 2-sample t tests (α = 0.05) were performed on all MS and HC groups. Bars indicate standard deviation, and asterisks indicate significant difference between groups (P < 0.05). Mean diffusivity values of all tracts in VD group were higher than those in voiders group. Voiding dysfunction versus voiders also showed significant difference approaching significant difference (P = 0.054) in the left ATR.

DISCUSSION

Several models have been used to assess the integrity of WMTs in neurogenic patients. This study is the first, to our knowledge, to use DTI to assess WMTs and their correlation to VD, specifically in women with MS. White matter tracts in MS patients are characterized by demyelinated axons, which limit the anisotropic diffusion of water in these fibers and, subsequently, limit signal transduction along the axons.²¹ In damaged tissues, FA is lower than normal because of loss of coherence in the main diffusion direction, and MD is higher as a result of increased free diffusion of the fiber tract.¹⁴ Our results verify that there is significantly lower FA and higher MD, indicating more damage in both WMTs of interest (right and left) in all MS patients with NLUTD compared with HCs. Within MS women, there was a trend of lower FA and higher MD in both ATR and SLF tracts on both sides in VD compared with spontaneous voiders, indicating damage in these tracts in MS women who had VD, although only MD values of the left ATR demonstrated difference approaching significance (P = 0.054). Interestingly, significant differences were detected between VD and HCs in both tracts, but not between voiders and HCs, suggesting that although the integrity of the 2 proposed tracts was compromised in MS women with VD, it was still preserved in MS women who void spontaneously as they were similar to HCs. In addition, more damage in the left and right AF—a WMT involved in language processing but not the bladder cycle,^11–13 in all MS women compared with HCs, but not between MS groups (voiders and VD), suggests that damage in the ATR and SLF is truly related to bladder control in MS patients and should not be attributed as global damage because of the advancement of the disease. These results further confirm the importance of the ATR and SLF tracts in normal function of the lower urinary tract, specifically the degradation of these WMTs in MS patients with VD.

The ATR refers to the fiber pathways that connect the anterior nuclear group and midline nuclear group of the thalamus to the medial prefrontal cortex (mPFC), an area that has been identified in numerous stroke and trauma studies as critical to long-term incontinence.^22,23 The SLF is considered an association fiber tract and connects the parietal cortical areas to different frontal cortical regions.²⁴ Results of this study suggest that both WMTs (on both sides) exhibit similar trends with regard to the loss of integrity in all MS women with NLUTD. These findings are expected as greater WMT damage associated with lower urinary tract function is related to more severe clinical manifestations. The mPFC is part of one of the 2 proposed supraspinal circuits that leads to the final decision to void. The decision is relayed down to the periaqueductal gray and pontine micturition center to switch the bladder from storage to void in healthy individuals.²⁵ Original positron emission tomography studies revealed that the dorsal and mPFC were activated in healthy women who could successfully void in the scanner.²⁶ Since then, multiple other studies have evaluated the initiation and maintenance of voiding in healthy individuals, and the results have supported the involvement of the frontal and prefrontal cortex.^27–31 Damage to the ATR and SLF, therefore, could result in weakened connectivity between the frontal and prefrontal regions, possibly contributing to VD in the MS patients.

Many studies have yielded results similar to our findings that both proposed tracts, especially the ATR, are important for proper bladder function. A study by Tadic et al³² analyzing fMRI signals during bladder filling in women who had urgency urinary incontinence revealed that there was more damage in patients who had detrusor overactivity. In another study, Tadic et al⁵ found a correlation between the ATR and SLF tracts in older women with urinary incontinence. Their study assessed the structural changes in the brain’s white matter hyperintensities (WMHs) and found that brain deactivations became more pronounced with increasing WMH burden. Increasing global WMH burden might reflect the presence of WMH in both the ATR and SLF, thereby affecting the connection between key regions involved in the micturition and voiding. Using different metrics—FA and MD values obtained from DTI brain scans—in our study, the importance of ATR and SLF tracts in proper bladder function is again reiterated, specifically in MS women.

Limitations

Most MS patients originally present with storage symptoms such as frequency, urgency, and urge urinary incontinence. As their disease progresses, patients will start experiencing voiding phase symptoms. There is no agreed classification for VD in women; thus, studying and managing VD become a challenge.^33,34 Although abnormally slow urine and high PVR are the basis for the diagnosis of VD (according to the International Continence Society and International Urogynecological Association³⁵), the normal ranges for these parameters themselves vary in literature. Voiding dysfunction specifically in MS women is a spectrum of disease with being in retention and catheterization dependent on one end and having hesitancy and incomplete bladder emptying on the other. Therefore, we decided to define MS women with VD as patients who perform self-catheterization or have a PVR volume of 40% or more MCC (instead of assigning an absolute value for PVR).

The inherently unnatural setting of lying supine for the fMRI and UDS is a confounding factor that may increase voiding difficulty especially in those who already have difficulty emptying their bladder. Therefore, patients were categorized to their corresponding groups based on their clinical presentation and PVR, not the values obtained during fMRI/UDS. Although supine positioning, UDS catheters, and supraphysiologic bladder filling may affect the activation and deactivation of different brain regions (for BOLD signal analysis), the DTI signals in the WMTs have not found to be altered by these quick dynamic variations in similar studies on lower urinary tract control.^36,37 In addition, the study was done in MS women only, where the disease is dynamic and heterogeneous in terms of lesions location, disease course, and level of disability, making it difficult to generalize observation. Our study evaluated differences in supraspinal activation during micturition; however, spinal cord lesions are also common in MS and have major effects on sensory and motor function. We controlled for ambulatory status, disease stability, and disability status but did not control for total lesion burden or location. Closer matching of subjects’ ages should also be considered in the future to eliminate it as a confounder. Finally, because the involvement of white matter in lower urinary tract control is still yet to be thoroughly investigated in the field, the power of this preliminary study and clinical meaning of these results still need to be further analyzed and confirmed with future studies.

Despite these limitations, this is the first study, to our knowledge, to examine the integrity of ATR and SLF in MS women with and without VD, showing that there is damage to the 2 proposed WMTs involved in lower urinary tract function in MS women, more specifically patients with VD in this cohort. Assessing the damage of these WMTs may be used as an independent predictive indicator for VD in MS patients. Further studies on larger populations of male and female patients who have VD with different types of neurologic or nonneurogenic voiding disorders should be done to investigate the sources and causes of damage in these WMTs and to gain deeper insight into how the supraspinal centers control bladder function and its relationship to VD.