Thesis Abstract

Objective The clinical entity “Symptomatic Tarlov Cysts” is a highly under reported condition. We aimed to perform an electrophysiologic evaluation in patients with Tarlov cysts to determine whether the cysts create electrical abnormalities that could translate into clinical symptoms. The findings are correlated with the data currently available in the literature.
Methods Thirty patients with unexplained pelvic, sacral, perineal and/or leg pain who harbored small and/or large Tarlov cysts were selected at an outpatient clinic for physical medicine in musculoskeletal disorders. An MRI of the lumbosacral spine of each patient was reviewed. An experienced physiatrist acquired information related to pain and paresthesia in addition to bladder, bowel and sphincter symptoms. An expert electrophysiologist performed nerve conduction and electromyography studies on the patient’s legs and the pelvic floor.
A review of the case reports on Tarlov cysts was performed. The symptoms of the patients in the study were compared with the symptoms reported in reviews and case reports.
Results In all cases, the presence of Tarlov cysts was associated with sensory neuron symptoms, such as pain and paresthesia, and with bladder, bowel, sexual, and/or sphincter complaints. In all cases, electromyography documented axonal damage in multiple lumbar and sacral nerve root myotomes.
Conclusion Symptomatic Tarlov cysts clinically and electrophysiologically represent a progressive chronic cauda equine syndrome. In patients with intractable sacral, perineal, pelvic or leg pain, symptomatic Tarlov cysts should be included in the differential diagnosis

Thesis Overview

Tarlov cysts (TCs), or perineural cysts, are spinal meningeal cysts that contain nerve root fibers. They originate from dilations of the nerve root sheath at the dorsal root ganglion and are most frequently identified in the sacrum [1,2]. In 1938, when Isadore Tarlov discovered sacral cysts, he first believed that they were benign incidental findings. However, he later reported that they could be a source of radicular pain [3,4]. Unfortunately, this first erroneous assumption perpetuated, resulting in a worldwide consensus regarding their clinical insignificance [5]. There are also few clinical diagnostic characteristics that can be used to link these cysts to specific symptoms. Hence, it is almost always concluded that a patients’ pain originates from other coexisting conditions, primarily degenerative disorders that are observed on imaging [6]. Many physicians are therefore unaware that -sacral symptomatic Tarlov cysts” (STCs) are a well-defined clinical entity that includes debilitating pelvic, perineal and leg pain and urogenital and bowel disturbances. STCs are easily identified when a comprehensive history is obtained [2,5,7,8]. Because this pathology is systematically overlooked, several authors have suggested that the incidence of STCs is significantly underestimated [2,5,8-16]. The purpose of this paper is to report electrophysiological findings in patients with unexplained chronic pelvic, perineal, and leg pain syndromes who harbor lumbosacral TCs to determine whether TCs create electrophysiological abnormalities that could translate into clinical symptoms. We used nerve conduction studies (NCS) and needle-electromyography (EMG) to analyze the lumbosacral nerve root myotomes. To determine the pathogenic activities and mechanisms that may lead to the various symptoms of this debilitating disease, our findings from NCS/EMG were linked to the information available regarding STCs in the literature.

Methods

The electronic database of a physiatrist outpatient clinic for patients consulting for musculoskeletal pain was searched to identify patients with unexplained debilitating or refractory lower back, pelvic, perineal and/or leg pain. An MRI of the lumbosacral spine, the sacrum and/or the pelvis was reviewed to determine whether TCs were present. TCs with a minimal size of 5 mm (or 3 mm for those that were located on the smaller lower nerve roots (S3-S4)) were analyzed [16]. If TCs were detected, the patients were specifically questioned regarding pain and numbness in the leg, perineum, bladder, bowel or sphincter dysfunction, genital symptoms, neck pain and headaches [2]. The painful myotomes in the leg (L3 to S2) and perineum (S3-S4) were recorded. Myotomes S1 and S2 were considered together and defined as pain on the posterior side of the leg because patients usually reported a diffuse distribution. The first 30 patients found to harbor TCs (27 women and 3 men) were selected for further investigation. NCS and needle EMG were performed by an expert senior neurophysiologist. The neurophysiologist was informed of the location of only the largest TC in each patient but was blinded to the MRI images. The electrodiagnostic tests consisted of conduction studies performed on the sensory sural nerves (which contain fibers from nerve roots S1 and S2), the motor peroneal nerves and the S1 Hoffmann reflexes (the electrophysiological equivalent of the Achilles tendon reflex). The needle EMG included the L3 to S3-S4 myotomes: L3 (vastus medialis muscle); L4 (vastus lateralis muscle); L5 (extensor digitorum muscle); L4-L5 (tensor fascia lata muscle and tibials anterior muscle); S1 (gastrocnemius muscle medial head); S2 (tibial nerveinnervated intrinsic foot muscles) and the S3-S4 myotomes (external anal sphincter). The diagnosis of a S1 radiculopathy was based on abnormalities during needle EMG and not solely on the S1 Hoffmannreflex latency. Additionally, the S3-S4 ano-anal reflex (the electrophysiological equivalent of the ano-cutaneous reflex) was analyzed. This reflex is meant to prevent fecal incontinence, and it therefore requires an intact reflex arc that consists of both afferent and efferent limbs. When using needle-EMG on the anal sphincter, only the efferent (motor) limbs were evaluated, whereas when analyzing the ano-anal reflex, the afferent (sensory) limbs were also assessed. This perineal reflex also indirectly provided information about the urethral sphincter [17]. During needle EMG, the presence of denervation potentials or the presence of at least 50% polyphasic Motor Unit Potentials was considered abnormal. Most muscles normally show only 5-25% polyphasic MUPs (e.g. about 5% in the biceps muscle and 25% in the extensor muscles of the upper and the lower limbs). The muscle with the highest percentage polyphasic MUPs in most individuals is the adductor longus muscle with sometimes up to 40% of polyphasic MUPs. Therefore, we used ≥ 50% polyphasic MUPs as the absolute cut off value, so as to be sure that we didn’t overcall the findings. In our study, patients with sacral TCs often complained about neck pain. A MRI of the cervicodorsal spine was therefore reviewed, if available. Because the headaches that were often reported in our patient group might have reflected an increase in pressure inside the cerebrospinal canal [2,18], a complete ophthalmologic examination that included optic disc biometry was performed by an expert ophthalmologist. The examination used optical coherence tomography (Spectralis OCT, Heidelberg) and static automated perimetry (Humphrey Field Analyzer, Zeiss or Octopus 300, Heig-Streit) to detect and grade possible papilledema. To explore STC pathogenesis and identify the typical patterns of symptoms that are caused by STCs, the literature from 1955 through 2016 was searched to obtain relevant reviews and case reports about patients with TCs. The MESH terms Tarlov, perineural, sacral and/or meningeal cysts were used. To gain new insights, the results of the NCS/ EMG tests were correlated with the data obtained from the literature.

Results

The characteristics and symptoms of the patients are shown in Table 1. The ages of the patients ranged from 25-74 years old (mean age, 46.0 ± 11.8 y). The size of the largest TC in each patient varied from 3 mm to 36 mm (mean 7.9 ± 3.8 mm). Most cysts were located on the L5 to S4 nerve roots. The patients were symptomatic for between 8 months and 50 years (mean 11.6 ± 12.0 y). Table 1: Characteristics and symptoms of patients in the case reports and this study. The EMG results are listed in Table 2. NCS revealed sural nerve abnormalities in 5 (16.7%) of the patients and increased latency in the S1 Hoffmann reflexes in 6 (23.1%) of the patients. Of the 28 patients who underwent ano-anal reflex testing, 25 (89.3%) displayed increased latency on one or both sides. None of the patients showed motor peroneal nerve abnormalities. Table 2: Results of NCS and needle EMG studies showing the mean number of cysts and the mean size (range) of cysts on each nerve root (Reference values: sural nerve amplitude ≤ 5 μV and/or latency ≥ 3.5 ms at 14 cm [53], S1 Hoffmann reflexes ≤ 30 ms, ≤ 32 ms and ≤ 34 ms for body length ≤160 cm, ≤182 cm and ≤195 cm repectively or a left/right difference of 2.0 ms; ano-anal reflex latency ≤ 50 ms or left-right difference ≥ 5 ms, or amplitude of L-R difference ≥ 500 μV [57]). Following needle EMG, abnormalities were identified in multiple lumbosacral nerve root myotomes in all of the patients, and these abnormalities were bilateral in 80% of the patients. In 25 (83%) patients, the pain was located in two or more dermatomes at the same time. Only 17 (57%) patients predominantly complained about pain in one leg. Ten (59%) of these 17 patients had the largest TC on the same side as the most severe pain. In 15 of the 17 (88%) patients, the most severe EMG abnormalities were found in the most painful leg. In all the patients, TC and/or nerve root dilations were found on all of the nerve roots of the painful dermatomes. Three patients with symptomatic sacral TCs and EMG abnormalities in the sacral nerve root myotomes also had severe thoracic pain in a dermatomal pattern. Therefore, an EMG of the thoracic paraspinals was performed, and in these cases, EMG abnormalities were detected in the corresponding dermatomes. The MRIs showed nerve root dilations in the corresponding nerves, and the nerve root fibers were splayed out. Additionally, 23 (76.6%) patients also complained about neck pain and/or pain in the arms. Their MRIs of the cervicodorsal spine were therefore revised. A previous MRI was available in 18 (60.0%) patients. In 15 (88.2%) of these patients, nerve root dilations were identified in the lower cervical and/or upper dorsal nerve roots (C7 to Th4). Fundoscopy and OCT showed no papilledema, and a visual field examination showed no abnormalities in any of the 30 patients.

Discussion

From the literature search, a total of 79 case reports and reviews describing 507 cases (236 women and 100 men) were identified, and the results of these studies were compared to those in our cohort. In the literature, only rare cases of large TCs that cause significant neurologic symptoms are generally reported because they are obvious and cannot be easily overlooked. The results of our study, in combination with a more thorough analysis of the case study-based literature and reviews, demonstrate that even small sacral TCs can cause debilitating sacroiliac, pelvic and genital pain, sexual dysfunction, bladder and bowel symptoms, and urinary and fecal incontinence. Pathogenesis Increased intraspinal pressure: TCs emerge as a result of fluid becoming trapped inside a nerve root. The work of Sun et al., who microscopically analysed resected TCs, increased our understanding of the mechanisms underlying cyst formation. The upright position of humans causes the physiological hydrostatic pressure in the terminal dural sac to be high. Under increased pressure, the lumbar and sacral nerve root sheaths, which contain the nerve root fibers that leave the terminal dural sac, may begin to enlarge. Eventually, for unknown reasons, fluid may become trapped between the endoneurium and the perineurium, and some of these cases may evolve into cyst-like structures: the so-called perineural cysts or TCs [19]. This expansion of the nerve root sleeve may be enhanced by congenital weakness of the connective tissue because large TCs have been found to occur more frequently in patients with hypermobility and connective tissue disorders, such as Marfan syndrome [2,20]. Sun et al. [19] compared the mechanisms involved in cyst initiation and enlargement with the formation of aneurysms in arteries. In aneurysms, years of increased pressure force a weakened vessel wall to enlarge. Similarly, nerve root dilation may require a long period of increased hydrostatic pressure in the spinal canal [2,19,21-24]. A ball-valve mechanism has been described in some original reports and reviews on TCs. This valve mechanism ensures that when CSF comes under increased hydrostatic pressure, it enters but does not leave the cyst. This allows pressure to build up inside the cyst. On CT myelography, cysts with a valve mechanism do not immediately fill up with contrast medium. Where the root passes through the foramen, the presence of a valve mechanism can be peroperatively confirmed using the Valsalva maneuver [4,6,25,26]. The valve mechanism theory does not, however, cover all cases because several studies have reported evidence demonstrating that non-valved cysts can also cause symptoms [2,6]. Additionally, it is well established that aseptic meningitis can occur when treating STCs with aspiration and fibrin glue injection if the glue is injected into a non-valved cyst from which it can spread into the dural sac [12,27]. Potts et al. [28] proposed that in cases with communicating cysts, the pressure within the cysts should be equal to the pressure in the subarachnoid space. The literature therefore indicates that TCs without a valve mechanism are capable of producing significant symptoms. However, the mechanisms underlying these symptoms remain unclear. Using CT-myelography, we documented both valved and nonvalved STCs in one patient in our cohort (Figure 1). The MRI of this patient showed that there was a large cyst (36 mm) on nerve root S2 that caused significant bone erosion (Figure 1A). Smaller TCs were observed on the bilateral L5, bilateral S1, and right S2 nerve roots (sizes: 8 mm-14 mm) (Figure 1B). Neither of the S2 cysts immediately filled on CT myelography, and they were therefore categorized as valved cysts. In contrast, the smaller cysts that filled immediately were categorized as non-valved cysts (Figure 1C). This patient showed bilateral atrophy in the S2-supplied intrinsic foot muscles in addition to significant bilateral EMG abnormalities in these S2 myotomes, which contained the valved cysts. However, she also developed a partial left foot drop (extension strength 3/5) with significant EMG abnormalities in the left L5 nerve root-supplied myotomes, on which a smaller non-valved TC (8 mm) was located. This case demonstrated that larger valved and smaller nonvalved cysts (which are actually nerve root dilations) can be present in the same patient and can produce neurological symptoms as a result of axonal damage that can be demonstrated using EMG. Consequently, we conclude that in non-valved cysts, a prolonged increase in hydrostatic pressure in the spinal canal is the crucial factor that contributes to the irritation of the nerve root fibers inside the nerve root sheath. We hypothesize that under a high level of intraspinal pressure, some of the nerve root sleeves become merely dilated, whereas others form true cysts. This mechanism is indeed similar to the mechanism that leads to the formation of a vascular aneurysm [19] because despite a prolonged increase in arterial pressure, only a minority of hypertensive patients develop an aneurysm. Additional evidence has been reported to support the role of increased intraspinal pressure in patients with STCs. When performing spinal fluid taps, Feigenbaum and Henderson observed persistent increases in opening pressure in 1 out of 10 STCs. They also suggested that increased intracranial pressure may be associated with STCs [2]. Anatomical pathological evidence comes from studies that have analyzed resected nerves that were obtained from autopsies. Large TCs were found caudally on the lumbar and sacral roots, whereas smaller cysts were located on the thoracic nerve roots. These data suggest that in more cranial locations, intraspinal pressure may be elevated at least high enough to facilitate the formation of TCs [21]. .These data are in accordance with the findings reported in our study. Twenty-three patients with sacral TCs complained of significant neck and/or arm pain. A MRI of the cervical spine was available in 18 of the patients, and 15 of these patients also harbored multiple nerve root dilations or smaller TCs (<10 mm) in the cervicodorsal region. An example of this type of nerve root dilation that was observed in the patient who had large sacral TCs is shown in Figure 1 (Figure 1E). Moreover, three patients with unilateral thoracic pain also demonstrated thoracic radiculopathy on needle EMG. They showed no obvious TCs, but the nerve fibers were splayed out inside the nerve sheath. Several authors have described that patients may report experiencing more pain while in an upright position, such as walking, standing or sitting, or when performing Valsalva maneuvers. This pain is often relieved when the patient is lying down, indicating that it may be affected by the pressure inside the spinal canal [2,10,19,29]. In our study, 30% of the patients reported increased pain while performing Valsalva maneuvers, while 86.7% reported increased pain while standing and 83.3% reported increased pain while straining. In 66.7 % of the patients, this pain was immediately relieved when the patient lied down. However, the majority (86.6%) of the patients could not lie down for long periods of time because of local pain on pressure points (e.g., the sacrum and the greater trochanter). In women, the symptoms may be exacerbated during pregnancy [30]. Among our patients, 12 out of the 16 women who had been pregnant reported pelvic or lower back pain during pregnancy. Additional evidence of increased intraspinal pressure indicates a higher risk of experiencing a dural leak while performing surgery for STCs [6,11,19,25,31,32]. This is because of the fragility of the tissues and the elevated pressure in the spinal canal [10]. Finally, a high frequency of headaches might also reflect increased pressure in the cerebrospinal canal [1,2,8,11,18,33]. Increased intracranial pressure: It is likely that there is a link between STCs and idiopathic intracranial hypertension (IICH) because a combination of symptoms, including headache with pain in the neck, the back or the limbs, has been reported in patients with benign IICH. This condition occurs primarily in young obese women and is associated with papilledema and visual problems. Bortoluzzi et al. [34] discovered that in these patients, the spinal root sleeves were markedly dilated. Additionally, several other authors have reported radiculopathy in patients with IICH [35-38]. Obeid et al. observed radiculopathy during a needle EMG of the paraspinal muscles in a patient with IICH [39]. In a study of 101 adults with IICH, Round and Keane reported observing neck stiffness in 31, paresthesia in 22, and lower back pain in 5 patients. They attributed all of these symptoms to spinal nerve root irritation because all of the symptoms resolved immediately following lumbar puncture [40]. In a recent prospective study by Wall et al. that included 165 patients with IICH, 84% of the patients reported headaches, and half of the patients (53%) had back pain, while 42% had neck pain and 19% had radicular pain [41]. Obesity is associated with increased intra-abdominal pressure, which secondarily increases cerebrospinal pressure to a level higher than that observed in lean patients. One of our patients who lost 20 kg of body weight over 3 years reported marked improvement of pelvic pain and in the intensity and the frequency of headaches. Of the cases reported in the literature, only 8 (1.6%) were reported to have headaches, and this is most likely because the link between STCs and headache remains unclear. The presence of headaches was therefore not investigated. Thus, we propose that clinical entity -STCs” may essentially correspond to a -chronic idiopathic moderate cerebrospinal hypertension syndrome”. To investigate the above hypothesis, we screened our patients for papilledema. We did not observe papilledema in any of our patients. This is probably because the pressure inside the cerebrospinal canal in STC patients is likely to be substantially lower than the pressure in IICH. Further investigating this hypothesis would require invasive craniospinal pressure monitoring [42]. Size and multiplicity: Multiple TCs can originate as a result of the above-described increases in intraspinal pressure. This process has been confirmed in many studies in the literature. Smith [21], who microscopically examined nerve roots from 100 autopsies, found cystlike formations in 9 patients. They were all multiple, and most were symmetrical. The number of cysts per patient ranged from 5 to 13. While most of the cysts were macroscopic in size and readily visible at autopsy, several could only be seen in microscopic preparations. Similarly, in his autopsy studies, Tarlov observed predominantly multiple sacral cysts [4,25,26]. Langdown et al. identified multiple and small TCs in perioperative investigations of the majority of patients with lumbosacral complaints. Because they were small, the authors assumed that the patient’s radicular pain was attributable to other lesions that were identified on MRI and that the small TCs were therefore clinically irrelevant [6]. However, it was not reported whether these patients suffered from perineal pain or urogenital and bowel problems. In a retrospective study, Komisaruk and Lee reviewed the MRIs of 18 women with persistent genital arousal disorder (PGAD) and found TCs in 12 (66.7%). The cysts had sizes that ranged from 3 mm to 20 mm (mean 9.6 ± 5.1 mm). The authors concluded that smaller cysts usually produce sensory symptoms [16]. This conclusion was confirmed in a more recent study by Sun et al. in which a negative correlation was found between the size of the cysts and their multiplicity. They also noted that when cysts are small, the pressure inside may still become critical. Therefore, even very small cysts can be symptomatic, and the initial symptoms in such cases are pain and paresthesia, which can be severely debilitating [19,24,43]. These data are in accordance with the findings in our study, which showed that all of the affected patients had either multiple smaller cysts or clear nerve root dilations and EMG abnormalities in multiple nerve root myotomes (Table 2) in addition to bilateral pain or paresthesia. The EMG abnormalities corresponded more often to the painful dermatome than the radiologic findings (Table 2). According to Naderi, fluctuations in pain severity and a lack of correlation between radiologic findings and symptomatology are the main characteristics of non-operated Tarlov cysts [44]. This result is in accordance with the findings of our study. The pain was usually not in one dermatome, but in several dermatomes at the same time. Moreover, the most severe pain was not always on the side of the largest cyst [8]. To date, the reason for this lack of correlation is not clear. These findings confirm that it is not the size of the cysts but rather the pressure inside the nerve roots that is responsible for these symptoms [6,19,45]. Figure 2A shows a MRI myelography that was obtained from a 64-year-old male in our cohort who suffered for almost 50 years from unexplained debilitating sacral and leg pain. The needle EMG demonstrated bilateral radiculopathy in multiple lumbosacral myotomes. These results clearly highlight the clinical relevance of small TCs, despite the fact that it remains widely assumed that small, multiple TCs are clinically irrelevant. Location Sacral lesions are usually located on the S2 and S3 nerve roots [4,7,23,46]. In our study, all patients harbored TCs on either the S2 or the S3 roots or both, and all patients had EMG abnormalities in the S2 or S3-S4 myotomes or both. Sun et al. reported that nerve root L5 is frequently involved [24]. In our study, TCs and EMG abnormalities were also frequently detected in the L5 myotomes. Although 96.7% of the patients had TCs on the S1 roots, only 40% had EMG abnormalities in the S1 myotomes; however, 81.7% had TCs on the L5 nerve and as much as 76.7% had EMG abnormalities in the L5 myotomes (Table 2). It is likely that the S1 root has more space to expand into the lateral recess than the L5 root has. This increase in available space may compensate for the increased pressure inside the root. It is worth noting that the L4 nerve root was also involved in 36.7% of our patients. These percentages of EMG-abnormalities in the L4 to S1 myotomes would be high in the population being referred for radiculopathy. However, this is an extraordinarily elective group of patients with a longstanding history of unexplained pain. In these patients, the EMG abnormalities have developed over many years. As in other neuropathies, a longer compression of the nerve fibers (axons) in the nerve root(s) is associated with a higher probability that more severe and more widespread EMG abnormalities will occur. Onset of symptoms The onset of symptoms usually occurs in patients in their 4th or 5th decade [23]. However, symptoms can appear in teenagers [45,47]. In our cohort, the first symptoms developed at 14-19 years old in 3 patients, in their 20s in 11 patients, in their thirties in 7 patients, in their forties in 4 patients, in their fifties in 4 patients, and at the age of 64 in 1 patient. The mean age of onset (34.4 ± 12.9 y) in our patient group that had smaller cysts (on average) was approximately one decade younger than the age of onset in patients with large TCs in the case studies reported in the literature (46.0 ± 11.8 y). However, in our study, the two patients with the largest cyst sizes (34 mm and 36 mm) were also in their fifties when they first developed symptoms, in accordance with the average age of onset reported for large cysts in the literature. Onset may occur when there is a history of trauma or heavy straining [30]. In our study, 16.7% of the patients reported that trauma was the trigger that elicited or aggravated their symptoms. The first symptoms are usually present long before an official diagnosis of STCs is established. In case reports in the literature, the time before diagnosis was 4.4 ± 6.7 y (range 1 week-25 y), whereas in our patient group, it was 11.6 ± 12 y (range 1 y-50 y). The apparently longer interval between initial symptoms and diagnosis in our study may have been because we inquired after less obvious sensory symptoms associated with smaller STCs. Prevalence Previous studies that analysed the MRIs of patients with lower back pain have reported a prevalence of 1.5%-4.6% [6,22,48,49]. However, because lower back and leg pain are almost always attributed to lumbar degenerative changes or disc problems, these MRIs did not routinely include axial or coronal sequences of the sacrum. Therefore, small TCs may have been overlooked. MRI studies of the lumbosacral spine that accounted for small cyst sizes have reported a prevalence of 9.1%-10.6 % [15,50]. When considering TCs that cause symptoms, it has been estimated that approximately one-fourth of such TCs are symptomatic at the time of discovery [6,22]. However, an unknown number of non-symptomatic cysts may also become symptomatic later in life [5]. TCs were found to be more prevalent in specific patient groups. For example, Tani et al. reviewed the pelvic MRIs of 102 women with otherwise unspecified gynecological problems and found that 10 (8.9%) had sacral meningeal cysts [46]. Likewise, Van de Kelft and Van Vyve [9] showed that of 17 patients with perineal pain, 13 had TCs (75%), while Komisaruk and Lee [16] found TCs in 12 out of 18 women (66.7%) with PGAD. It has been reported that 61.2%-87% of patients who harbor TCs are women [5,6,30,48,50]. We found that this number was 74.8% in case reports and 90% in our EMG study. The reason f