Cauda Equina Syndrome: presentation, outcome, and predictors with focus on micturition, defecation, and sexual dysfunction

Even though micturition, defecation, and sexual function are substantially affected in cauda equina syndrome (CES), data on outcome are scarce. Medical files of patients operated on lumbar herniated disc were screened for CES and retrospectively analyzed for baseline characteristics, outcome of micturition, defecation, and sexual function and possible predictors. Seventy-five CES patients (52% men) were included with a mean age of 44 years. L5–S1 was the most common affected level. Duration of CES complaints at presentation was, on average, 84 h (median 48 h). Prevalence of symptoms at presentation: sciatica (97%), altered sensation of the saddle area (93%), micturition dysfunction (92%), and defecation dysfunction (74%). Only 26 patients were asked about sexual dysfunction of whom 25 patients experienced dysfunction. Female gender was associated with more defecation dysfunction at presentation than male gender (OR 4.11; p = 0.039). All patients underwent decompressive surgery. Two post-operative follow-up (FU) moments took place after a mean of 75 h and 63 days. Outcomes at second FU moment: micturition dysfunction 48%, defecation dysfunction 42%, sexual dysfunction 53%, sciatica 48%, and altered sensation of the saddle area 57%. A shorter time to decompression was associated with more sciatica at FU 1 (p = 0.042) which effect had disappeared at FU 2. This study is unique in (1) displaying the presenting features in a large cohort of CES patients, (2) demonstrating that recovery after decompression is slow and far from complete in the majority of patients with regard to micturition, defecation, and sexual function and (3) evaluating predictors for outcome.


Introduction
Cauda equina syndrome (CES) is a neurological condition caused by compression of the cauda equina, most commonly described as a combination of sensory loss of the saddle area, motor deficit and/or loss of reflexes of the lower limbs, micturition dysfunction, defecation complaints, and/or sexual dysfunction [1,2]. The first article about CES appeared in 1934, in which a combination of neurological and urological complaints in patients with a ruptured intervertebral disc was described [3]. A herniated disc is still the most common cause of cauda equina compression; in literature, 45% of cases of CES are attributed to a lumbar herniated disc [1].
In the last decades, especially the topic of timing of decompression and its relation to outcome has gained much attention in literature, with several small studies showing better-albeit not always significant-outcomes after the early decompression [4][5][6][7][8][9]. Other studies could not demonstrate a better outcome after early decompression [10,11]. The value of urgent decompression was most convincingly showcased in the meta-analysis of Ahn et al., showing a better prognosis of sensory, motor, urinary, and rectal function in patients being decompressed within 48 h of presentation, compared to a group being decompressed after 48 h [12]. These results were confirmed by others [13,14].
In literature, there is a little focus on the prognosis of micturition, defecation, and sexual function [2]. This is remarkable considering the definition of CES. Recently, clinicians in spinal care were found to barely discuss sexual health and/or defecation at presentation and at follow-up, suggested to be due to, e.g., lack of knowledge or time [2,15,16]. Clinicians who do want to inform their patients about the prognosis of these functions are confronted with scarce data. This study was performed to (1) evaluate outcome of micturition, defecation, and sexual function in cauda equina syndrome after decompression and to (2) find possible predictors of outcome. In addition, presenting features of CES were analyzed.

Materials and methods
The medical records of patients operated in the Leiden University Medical Centre (LUMC; university hospital and referral centre for high complex spinal surgery) between January 1995 and September 2010 with the surgery code 'lumbar discectomy' or 'recurrent lumbar discectomy' were screened by two independent researchers (NSK, JAP) to identify patients with cauda equina syndrome (CES). Criteria to diagnose a patient with CES were, according to consensus of literature, one or more of the following: (1) dysfunction of micturition, defecation, and/or sexual function (not being attributable to use of opiates or previous disease), (2) altered sensation of the saddle area, with possible neurologic deficit in the lower limb (motor or sensory loss or reflex changes) [1,2]. Patients filed with a diagnosis of CES but not meeting those criteria were excluded. In case of doubt about the diagnosis of CES, a third assessor (CLAVL) was consulted. To check interobserver reliability between the two reviewers, 10% of cases were independently screened by both of them.
The following data were extracted from the medical file: (not derived through multiple imputation) are presented as corrected (corrected for the number of patients for whom documentation is available at the concerning follow-up moment) and as raw (not corrected for the number of patients). Due to an anticipated high amount of loss to follow up at FU 3, outcome at FU 2 was defined as main outcome and data at FU 3 were not used for regression analysis. For the patients for whom data at FU 3 are available, this will be mentioned separately.

Results
In the period January 1995-September 2010, a total of 744 surgeries coded as '(recurrent) lumbar discectomy' were performed at LUMC, for a total of 696 patients: 38 patients had surgery twice, 10 patients had triple surgery. Out of 696 patients, 75 patients (10.8%) were found to have CES. One female patient who underwent a lumbar discectomy twice, met CES criteria twice; however, since she had not recovered from her first CES completely, only the first surgery was included for analysis. Interobserver reliability regarding diagnosing CES was analyzed using Cohen's Kappa. There was substantial agreement between the two reviewers' judgement (j = 0.635).

Patient characteristics
For patients characteristics; see Table 1 and Fig. 1  whom 25 experienced sexual dysfunction. Documented problems were, e.g., altered sensation of genitals, inability to reach orgasm, erectile dysfunction, and priapism. For two patients, sexual status (active/nonactive) before onset of CES was documented; for the others, no notes on sexual activity were found. Anal sphincter tension and anal sphincter reflex (anal wink) were tested in 76.0 and 65.3% of patients, respectively, and were abnormal in the majority (63.2 and 59.1%, respectively). Abnormal anal sphincter tension was significantly associated with altered sensation of the saddle area (p = 0.007; Table 2), with a sensitivity for altered sensation of the saddle area of 68%. Abnormal anal sphincter reflex was not significantly associated with defecation dysfunction, although a trend was observed (p = 0.096; Table 3). Micturition dysfunction was not associated with either abnormal anal sphincter tension nor reflex (Table 4). Since data on sexual dysfunction at presentation were scarce, no analyses were done for sexual dysfunction.

Association patient characteristics: presenting features
The following factors were evaluated as predictors for presentation: age, gender, duration of complaints of herniated disc, duration of CES complaints, presence of altered sensation of the saddle area (for evaluating defecation and micturition dysfunction at presentation),    presence of micturition dysfunction (for evaluating defecation dysfunction and altered sensation of the saddle area at presentation), and the presence of defecation dysfunction (for evaluating micturition dysfunction and altered sensation of the saddle area at presentation). For sciatica at presentation, the regression model could not be run, nor could sciatica at presentation be added as a predictor to the other models due to quasi-separation of the data. Defecation dysfunction at presentation was significantly associated with female gender (OR 4.11; p = 0.039). Micturition dysfunction and altered sensation of the saddle at presentation displayed no predictors.

Surgery
All patients were decompressed by (partial) laminectomy and subsequent discectomy or sequesterectomy. A slight majority of patients (n = 36) was decompressed within 24 h after presentation (Fig. 3). Eight patients were decompressed more than 72 h after presentation at first doctor, with time to decompression of 96 h (n = 3), 120 h (n = 1), 138 h (n = 1), 168 h (n = 1), 192 h (n = 1), and 216 h (n = 1). In 7 of these cases, majority of the delay was caused by the first doctor (family doctor or neurologist) where the patient presented. Hereafter, surgery was performed within 24 h (n = 4), 48 h (n = 2) and 72 h (n = 1) after first presentation at the neurosurgeon. In one case, no discrimination could be made between delay at first and second doctor.

Post-operative outcome
The first follow-up moment (FU 1) at which micturition, defecation, sexual function, and/or altered sensation of the saddle area was documented was on average 75 h postoperatively (range 4-336 h; median 48 h). The latest time of FU 1 was 14 days post-operatively. Documentation on any of the items micturition, defecation, sexual function, and/or altered sensation of the saddle area at the second follow-up moment (FU 2) was available for 54 patients (72%), with a mean FU time of 63 days (range 4-300 days, median 60 days). A third follow-up moment (FU 3) at any item was documented for 23 patients (31%), with a mean FU time of 265 days (range 56-730 days, median 225 days). FU 3 is reported in the text as corrected, not pooled, and is not used in any regression analysis.

Micturition
Documented micturition dysfunction decreased significantly comparing pre-operative moment with FU 2 (92.0 versus 47.7%, p \ 0.001), Fig. 4. In one patient with postoperative dysfunction, requiring intermittent catheterization from the 4th day after surgery for a short period of time with complaints of urinary dysfunction up to the last follow-up moment, (6 months later), no complaints of micturition were documented at presentation. This was interpreted as misinformation at baseline. Pooled prevalence of micturition dysfunction was 47.7% at FU 2. Reported dysfunction included: indwelling catheter, onoff catheterisation, suprapubic catheter, reduced feeling of passing urine, and (mild) incontinence. At FU 3, 19 patients were evaluated of whom 11 displayed dysfunction (57.9%).

Defecation
Documented defecation dysfunction had decreased significantly after decompression measured at FU 2 (72.0 versus 41.8%, p = 0.004), Fig. 5. For three patients, defecation dysfunction was documented post-operative but not pre-op. Pooled prevalence of dysfunction was 41.8% at FU 2.

Sexual function
Due to scarce data on sexual function, investigating that proportions were not done; therefore, no p value for difference between pre-and post-operative dysfunction was derived. Corrected prevalence of sexual dysfunction at FU 2 was 53.3%, Fig. 6. Documentation on sexual dysfunction was done for 5 patients at FU 3; four of them displayed dysfunction (80%).

Altered sensation of the saddle area and sciatica
Documented altered sensation of the saddle area and sciatica both decreased significantly after decompression measured at FU 2 (93.3 versus 56.5% and 97.3 and 47.5%, respectively; both p \ 0.001), Figs. 7 and 8. At FU 3, 12 out of 18 patient with documentation reported altered sensation of the saddle area (66.7%), and 12 out of 20 reported sciatica (60%).

Predictors for outcome
The following factors were evaluated as predictors for outcome: age, gender, duration of complaints of herniated disc, duration of CES complaints, time to decompression (stratified groups, see Methods section for details of groups), and altered sensation of the saddle area/micturition dysfunction/defecation dysfunction/sciatica at presentation.
Due to quasi-separated data, sciatica at presentation had to be removed from all models except the one for micturition dysfunction at FU 2; altered sensation of the saddle at presentation had to be removed from all models except the one for sciatica at FU 1; micturition dysfunction at presentation had to be removed from the models for sciatica at FU 2 and altered sensation of the saddle area at FU 1 and FU 2.
Timing of decompression was found to be significantly associated with short-term outcome (FU 1) of sciatica: less time to decompression was associated with more sciatica at FU 1 (p = 0.042). After stratifying outcomes for time to decompression using the well-known break points from literature of 48 h [12,14] and 36 h [13], we found a statistically significant difference for outcome of sciatica at

Discussion
This is the largest single study performed about outcome of micturition, defecation, and sexual function in CES (n = 75). The incidence of CES among patients being operated on for herniated disc was relatively high in this study: 10.8% compared to 1-3% in literature [9,17]. This high incidence can be explained by the fact that the LUMC serves as a referral hospital for urgent neurosurgical cases. The CES definition that was used to include patients in this study is widely used in literature, and even though a univocal definition for CES does not exist, the authors believe that using this definition guaranteed a fair representation of CES patients. This study displays unique data on the presenting symptoms of a large group of CES patients, proving that, next to the well-acknowledged micturition dysfunction, also defecation and sexual dysfunction are common at presentation. Anal sphincter tension and anal sphincter reflex are often tested in CES patients, even though several studies found no diagnostic value for these tests [18][19][20][21]. In the current study, abnormal anal sphincter reflex was not significantly associated with any of the diagnostic criteria for CES used in this study (closest to an association was defecation dysfunction with p = 0.096). Abnormal anal sphincter tension is significantly associated with altered sensation of the saddle area (p = 0.007). Specificity of the test is 100% (all patients without altered sensation of the saddle area displayed normal sphincter tension) and sensitivity is 68% (a substantial proportion of patients with anamnestic altered sensation of the saddle area displayed normal sphincter tension). With a positive predictive value of 100% (all patients with abnormal sphincter tension had anamnestic altered sensation of the saddle area) and a negative predictive value of merely 19%, abnormal anal sphincter tension at physical examination supports the presence of altered sensation of the saddle area, but in no way rules out altered sensation of the saddle area, in case it is normal. Considering a specificity of 100%, the question arises whether it is necessary to test sphincter tension in patients without altered sensation of the saddle area, since it might cause unnecessary discomfort. Evaluating the anal sphincter reflex data, however, does indicate that sphincter testing might add extra information. Of the five patients with normal sensation of the saddle area at presentation, four were tested for anal sphincter tension (all normal). Two of those four were also tested for the anal sphincter reflex, which in one case was abnormal, demonstrating that with normal sensation of the saddle area and with normal sphincter tension, the anal sphincter reflex can still be abnormal. It could, therefore, be sensible to do anal sphincter reflex tests even in a patient with normal sensation of the saddle area. The fact that only a small proportion in our study sample demonstrated normal sensation of the saddle are at presentation (n = 5), limits our data in this aspect and, therefore, clinical relevance of sphincter tests cannot be secured nor refuted based on those findings.
Prevalence of micturition dysfunction is 47.7% of patients at FU 2. In an older study discussing 13 patients with CES due to herniated disc, the author states that ''in all such patients, there was incomplete return of normal micturition'' [22]. In a more recent study of McCarthy et al., a better recovery of micturition has been described: of 42 evaluated patients with CES due to herniated disc, 36% reported urinary incontinence (mean FU time of 60 months) [23]. The higher prevalence of micturition dysfunction in the current study might be due to the shorter FU time, assuming that function of micturition will improve gradually over time.
Regarding defecation dysfunction, this study found a prevalence of 41.8% at FU 2; McCarthy et al. found a higher prevalence at follow-up (60%): this could be due to attrition bias, the chance of which becomes greater when follow-up period of the study is longer-as in McCarthy's study. Sexual dysfunction in the current study was 53.3% at FU 2, which is quite similar to the prevalence of 57% reported by McCarthy et al. The true prevalence of sexual dysfunction might be higher than the ones found in studies; since the threshold to discuss sexual health is very high which, it is unlikely that either doctor or patient opens the topic, even if there are complaints.
Two patients in this study displayed a specific feature of sexual dysfunction: priapism. One patient reported spontaneous erections at presentation; unfortunately, no documentation on sexual function was done at follow-up for this patient. The second patient presented with a numb feeling of the penis which had changed to priapism at follow-up after 7 weeks and after 5.5 months. Priapism as a feature of cauda equina compression is extremely uncommon and only two reports in literature describe such a case: one involves a 61-year-old man with a herniated disc at L4-L5 [24], the other a 60-year-old man with a degenerative stenosis at L3-L4 and lumbar arachnoiditis [25]. Both patients experienced priapism and a numb, respectively, a burning sensation at the saddle area when walking, without sphincter disturbances. After decompression of the cauda equina, both patients experienced immediate and complete relieve of their symptoms, suggesting a causal relationship of cauda equina claudication and priapism. The parasympathetic fibers that are responsible for penile erection arise in S2-S4, and it is thought that their stimulation through (in these two cases: intermittent) compression had resulted in priapism [24]. To the authors' best knowledge, there are no case reports about priapism in nonintermittent cauda equina compression. Even though the course of priapism complaints of the two patients in this study is uncertain, the authors believe that it is not unthinkable that the priapism experienced by the two patients in this study might be attributed to compression of the cauda equina, even though exact mechanisms remain unclear.
This study found that female patients are more likely to present with defecation dysfunction than their male counterparts (OR 4.11; p = 0.039). Epidemiologic studies demonstrate that female gender is associated with more constipation, as a baseline fact in the general population as well as in the CES population when focusing on outcome [26,27].
Sciatica were more often present at FU 1 when time to decompression was shorter (OR 1.86; p = 0.042), which association had disappeared at FU 2. This association nowise undermines the importance of emergency decompression. It rather displays a correlation between duration of compression and other (prognostic) factors for which could not be corrected. Those factors are (1) the speed with which the compression has arisen, stating that patients with slowly developing anatomical lesions have a more favourable prognosis [28,29] and (2) the type of CES lesion, stating that a ''complete'' CES (with total obstruction) carries a poorer prognosis than a ''incomplete'' CES, adding that the incomplete type often occurs more gradually, (although not exclusively) [22,30]. It seems sensible that a patient with sudden onset of heavy symptoms in general presents earlier. This results in a shorter duration of CES complaints (patient delay) as well as a shorter time to decompression (doctor delay), even though outcome is poor (due to the extent of the lesion). Patients that show reasonable outcome when decompression is delayed by weeks [31], probably had a favourable anatomical lesion and more gradual onset of complaints. Since it is not (yet) possible to distinguish the group with the favourable conditions from the group with the less favourable conditions at the time of presentation (since, e.g., the exact correlation of clinical presentation and degree of canal obstruction on imaging is not yet known), it is necessary to decompress every CES patient as soon as possible, to create the best chances for fair recovery.
The authors strongly believe that this study's finding, e.g., that time to decompression is not associated with outcome of micturition, defecation, and altered sensation of the saddle area, and does display a correlation of more sciatica at FU1 when it is shorter, does not implicate that decompression in CES is to be delayed. First, the number of patients with delayed decompression in this study is relatively small (eight patients in the group decompressed after 72 h; five patients in the group decompressed within 48-72 h). Previously, meta-analyses have displayed better outcomes with decompression taking place within 48 h [12,14] or within 36 h [13]. Some studies with smaller patient numbers display a significant better outcome after earlier decompression as well, with significant better outcomes with decompression within 48 h [9] and even after 10 days [4]. Studies displaying no difference in outcome are a minority and evaluate relative small patient numbers only [10,11]. The finding in this study that a shorter time to decompression is associated with more sciatica at FU 1, should, therefore, be weighed by the small patient number of this study and the fact that this finding is not present at FU 2 anymore. In addition, in this study, clinical motivesunknown to the authors due to the retrospective study design-could have led to the decision for very early decompression in specific patient groups, which might have caused selection bias. The outcome would then be more influenced by factors on which clinical motives are based (and which have led to an early time to decompression) than by the actual time to decompression.
Significant predictors, such as duration, of CES complaints for more than 48 h as a risk factor for micturition dysfunction [32] and defecation dysfunction at presentation as a risk factor for sexual dysfunction [23] could not be identified in this study.
Missing data in this study are partly attributed to the inclusion of patients that were referred for surgery to LUMC, and were seen for follow-up at their original referring hospital (in which case the researchers did not have access to the follow-up data). Of the 19 patients that originated from either LUMC or the general practitioner, 100% displayed data at FU 2. Of the 55 patients that originated from a different hospital and were referred to LUMC for surgery, 34 patients (61.8%) displayed data at FU 2. The amount of dysfunction between the group of LUMC patients and the group originating from a different hospital and referred to LUMC was not significantly different at FU 2: neither for micturition dysfunction (42.1 versus 48.3%; p = 0.771) nor for defecation dysfunction (23.5 versus 45.5%; p = 0.193), sciatica (57.9 versus 46.9%; p = 0.565), or altered sensation of the saddle area (58.8 versus 63.6%; p = 0.767). This reason for missing data was, therefore, not believed to have caused bias.
Apart from loss to follow up of patients belonging to a different original hospital, three other reasons for missing data in this retrospective design are: (1) the topic was not discussed and, therefore, no notes are available, (2) the topic was discussed, but no notes were taken, and (3) the patient did not show up at FU moment. Especially for the latter two reasons, data are more likely to be missing when there are no complaints. To explore the extent of this bias-i.e., the hypothesis that data of patients with no complaints are more prone to be missing-patients with and patients without documented complaints at the previous check-up were analyzed for the amount of available data (Figs. 9, 10, 11, 12, 13). As is displayed, patients without documented complaints at their previous check-up are more prone to have missing data (apart from the case of defecation dysfunction, which cannot be easily explained). Due to this fact, the authors have chosen to not use FU 3 as main outcome parameter, even more since FU 3 is more likely to be planned for patients with complaints, therefore, contributing substantially to attrition bias when it would be used as a main outcome parameter.
Data on sexual function in this study were particularly scarce, as well at presentation as during follow-up, which resulted in the inability of performing several analyses regarding sexual dysfunction. Limited data on sexual function are ubiquitous in CES patients in literature [2], most likely due to barriers on both the patients' side as well as on the doctor's side, which could unfortunately not be minimized in this retrospective study design. The nature of Fig. 9 Availability of data on micturition. The proportion of available data per FU moment is stratified by patients for whom complaints were documented at the previous FU moment (black bar) and by patients for whom it was documented that there were no complaints at the previous FU moment (grey bar) Fig. 10 Availability of data on defecation. The proportion of available data per FU moment is stratified by patients for whom complaints were documented at the previous FU moment (black bar) and by patients for whom it was documented that there were no complaints at the previous FU moment (grey bar) Fig. 11 Availability of data on sexual function. The proportion of available data per FU moment is shown for patients for whom complaints were documented at the previous FU moment (black bar). For all patients for whom it was documented that there were no complaints at the previous FU moment, there was no data available at the next FU moment; therefore, there are no grey bars Fig. 12 Availability of data on sensation of the saddle area. The proportion of available data per FU moment is stratified by patients for whom complaints were documented at the previous FU moment (black bar) and by patients for whom it was documented that there were no complaints at the previous FU moment (grey bar) Fig. 13 Availability of data on sciatica. The proportion of available data per FU moment is stratified by patients for whom complaints were documented at the previous FU moment (black bar) and by patients for whom it was documented that there were no complaints at the previous FU moment (grey bar). For all patients for whom it was documented that there were no complaints at presentation, there were no data available at FU 1; therefore, there is no grey bar at FU 1 the available data on sexual dysfunction at presentation is striking: 25 out of 26 patients experienced sexual dysfunction. The fact that sexual function is more often documented for male patients than for their female counterparts is something that is believed by the authors to be due to both patient factors as well as doctor factors [15].

Conclusion
This is the largest single study about outcome in CES after decompression. It displays unique data about the outcome of micturition, defecation, sexual function, sciatica, and altered sensation of the saddle area, as well as presenting features of a large cohort of CES patients, proving that micturition, defecation, and sexual dysfunction are common at both presentation and at follow-up. Female patients were found to have significant more defecation dysfunction at presentation than their male counterparts. A shorter time to decompression was a risk factor for sciatica shortly after surgery (at follow-up moment 1), but not for long-term outcome. Studies on correlation between imaging at presentation and outcome could help identifying the patients being more at risk at presentation. Missing data were handled with multiple imputation with analysis of possible bias. A follow-up study is recommended for more longterm follow-up data. In conclusion, recovery after decompression for CES does take a long time and is not complete in a substantial amount of cases; something for which we should adequately prepare our patients when diagnosing CES.