Assessment of Female Patients With Rectal Intussusception and Prolapse: Is This a Progressive Spectrum of Disease?

Hotouras, Alexander M.R.C.S.; Murphy, Jamie Ph.D.; Boyle, Derek J. F.R.C.S.; Allison, Marion M.Sc.; Williams, Norman S. M.S.; Chan, Christopher L. F.R.C.S., Ph.D.

Abstract

BACKGROUND: Rectal intussusception may be the initial abnormality of a progressive pelvic floor disorder culminating in external prolapse. The evidence, however, is unclear, and the pathophysiological mechanisms underlying this condition are unknown.

OBJECTIVE: The aim of this study is to identify the relationship between age, symptom duration, anorectal physiology parameters, and type of intussusception/prolapse in order to appreciate the natural history of the condition.

DESIGN: All female patients diagnosed proctographically with rectorectal/rectoanal intussusception or external prolapse between 1994 and 2007 were studied. Demographics, symptom duration, and anorectal physiology results were compared between these proctographic groups. Patients with repeat proctographic evaluation were also analyzed separately.

SETTINGS: This investigation was conducted at a tertiary academic colorectal center.

PATIENTS: A total of 1014 women (median age, 51; range, 16–96), including 32 who underwent repeat proctography, were analyzed.

MAIN OUTCOME MEASURES: The primary outcomes measured were the differences in median age, symptom duration, and anorectal physiology parameters between the proctographic groups.

RESULTS: The cohort exhibited a statistically significant difference (p = 0.0001) in the median age of the proctographic groups with older patients diagnosed with rectoanal rather than rectorectal intussusception, which was supported by uni- and multivariate modeling. Symptom duration was statistically different (p = 0.0002) between the rectorectal intussusception (60 months; range, 1–936) and external rectal prolapse patient groups (36 months; range, 2–732). Patients with external rectal prolapse had statistically lower anal resting (median, 41 versus 77 cmH₂O) and squeeze pressures (median 40 versus 56 cmH₂O) than patients with rectorectal intussusception. Within 2 years, 19.2% and 3.8% of patients with rectorectal intussusception on the initial proctogram demonstrated progression to rectoanal intussusception and external prolapse.

LIMITATIONS: This study was limited by its retrospective nature.

CONCLUSION: Rectal intussusception may be an initial abnormality leading to external prolapse, but this appears to happen infrequently. Long-term observational studies are required to fully understand its natural history.
 

Rectal intussusception (RI) is defined as a mucosal or full-thickness infolding of the rectal wall that occurs during defecation.¹ The rectal wall may descend by varying degrees into the rectum (rectorectal intussusception, RRI), anal canal (rectoanal intussusception, RAI), or outside the anal verge resulting in external rectal prolapse (ERP). Patients with RI (RRI or RAI) and ERP frequently present with pain, obstructed defecation, or fecal incontinence. These symptoms may be attributed to 1) the presence of intussusception in the distal rectum and anal canal, 2) blockage of the rectal ampulla during defecation with a persistent desire to defecate, or 3) anal canal dysfunction secondary to an overt rectal prolapse.² Women are more commonly affected than men mainly because of the damaging effect of pregnancy/vaginal delivery on the pelvic floor and anal sphincter complex, in combination with the deleterious effect of the menopause.²,³

The exact pathophysiology of RI and ERP remains unclear, but loose sacral fixation of the rectum associated with an underlying connective tissue disorder or colonic dysmotility resulting in chronic straining are processes that are thought to be involved.⁴,⁵ In addition, the effects of pelvic floor dysfunction may be crucial to the development of ERP, with several theories suggesting a central role in this process for a wide variety of functional abnormalities, including increased intra-abdominal pressure and chronic straining during defecation,⁶ congenital abnormalities (eg, myofascial or connective tissue disorder), traumatic (eg, vaginal delivery) impairment of the levator ani plate with excessive distensibility (“ballooning”) of the urogenital hiatus, and perineal hypermobility.⁴,^7–10

Despite the plethora of theories addressing the cause of ERP, the natural history of this condition remains unclear. Studies have suggested that RI may be a progressive disorder that starts as a minor abnormality before eventually progressing to ERP.¹¹ This hypothesis is not confirmed, and, in fact, long-term observational studies demonstrate that progression of RI to ERP is rarely seen.¹²,¹³ Consequently, the mechanisms that initiate ERP remain elusive. If the hypothesis that ERP is the end point of RRI first evolving into RAI is correct, then a progressive relationship may exist between age, duration of symptoms, and the extent of RI/ERP. Therefore, the aim of our study was to explore the link between age, type of intussusception/prolapse, and duration of symptoms within an exclusively female cohort of symptomatic patients in an attempt to appreciate whether RRI/RAI and ERP are 2 distinct clinical entities or part of the spectrum of a progressive disorder. The secondary objective was to identify any anorectal physiological parameters that might shed light on the mechanism of development of RI and ERP.

METHODS

Clinical and Physiological Assessment

All female patients (n = 3686) referred to a GI physiology unit over a period of 13 years (1994–2007) with symptoms of fecal incontinence, obstructive defecation or both, underwent routine anorectal physiology evaluation including the measurement of maximum anal resting and squeeze pressures by the use of water-perfused anal canal pull-through manometry, sphincter morphology by endoanal ultrasound, pudendal nerve motor latencies using the St Mark’s electrode, and rectal sensory thresholds (eg, maximum tolerable volume, MTV) to air-filled balloon distension. All patients also underwent standardized evacuation proctography,¹⁴ and the presence of intussusception/prolapse was classified according to the Shorvon grading system (Table 1).¹⁵

 

TABLE 1. The Shorvon grading system for mucosal intussusception and prolapse

Female patients diagnosed proctographically with RRI (Shorvon grading 3–4), RAI (Shorvon grading 5–6), or ERP (Shorvon grading 7) were identified from our institutional GI physiology database. Women with a proctographic diagnosis of rectal intussusception graded 1 to 2 by the Shorvon classification were excluded from the analysis because of the uncertainty surrounding the clinical significance of these findings that are commonly seen in asymptomatic volunteers.^15–17 Data were reviewed retrospectively and included patient demographics (ie, age at presentation, sex), relevant medical history (ie, diabetes or neurological disorders), obstetric history (ie, number and type delivery) surgical history (ie, abdominal, pelvic, and perianal surgery) duration of symptoms of obstructed defecation or fecal incontinence and anorectal physiology measurements (eg, sphincter morphology, resting and squeeze pressures). Patients who underwent repeat proctographic evaluation as part of their clinical management had the findings of the initial assessment included in the main analysis; however, the initial and final proctograms were also examined separately for evidence of proctographic progression.

Statistical Analysis

The median age of female patients in each proctographic group (RRI, RAI, and ERP) was calculated, and intergroup comparisons were performed. Similarly, anorectal physiology findings (eg, sphincter integrity, anal sphincter resting/squeeze pressures, percentage of patients with rectal hyposensitivity (defined by MTV> 290mL)),¹⁸ and duration of symptoms of obstructed defecation or fecal incontinence were analyzed for each proctographic group, and intergroup comparisons were again performed.

Data analyses were performed by using commercially available statistical analysis software (GraphPad Prism, Version 5, GraphPad Software Inc, La Jolla, CA; SPSS v20, IBM, Armonk, NY). Data normality was assessed by using the De Agostino-Pearson omnibus normality test. Nonparametric data were compared by the use of the Mann-Whitney U test (MWU) where appropriate. Analyses of variance (using the Kruskal-Wallis test with the Dunn multiple comparison posttest) were performed when comparing median age, duration of symptoms of obstructed defecation or fecal incontinence, and anorectal physiology parameters between the proctographic groups (RRI, RAI, and ERP).

Both univariate and multivariate regression analyses were performed to examine the effects of a number of variables upon the development of RRI, RAI, and ERP. Logistic regression models were used to determine the association between the proctographic groups mentioned above and the following binary outcome measures: abdominal hysterectomy, abdominal prolapse surgery, anterior resection, caesarean delivery only, colposuspension, diabetes mellitus, episiotomy, forceps delivery, hemorrhoidectomy, IBD, neurological diseases, nulliparous, obstetric tear, perineal prolapse surgery, postanal repair, pelvic radiotherapy, rectocele surgery, sphincterotomy, sphincter repair, surgery for fistula-in-ano, surgical repair of anorectal atresia, vaginal hysterectomy, and ventouse delivery only.

For all analyses, a p value of <0.05 was considered statistically significant.

RESULTS

Demographics

A total of 1014 female patients (median age 51; range, 16–96) were included in the analysis. The cohort consisted of 826 patients (81.5%) with a history of previous vaginal delivery (median parity, 2; range, 1–10) and 188 (18.5%) who were nulliparous. Female patients with comorbidities (n = 628) included those with anterior/middle compartment prolapse (n = 54), diabetes mellitus (n = 54), neurological disorders such as multiple sclerosis (n = 10), pelvic radiotherapy (n = 7), and history of surgery (n = 33 abdominal/perineal rectopexy; n = 12 anal surgery; n= 3 anterior resection; n = 58 cholecystectomy; n = 337 hysterectomy; n = 22 rectocele repair; and n = 38 spinal surgery). A total of 386 women did not have any relevant identifiable comorbidities including 84 nulliparous women. The demographic and physiological characteristics of all female patients diagnosed with RRI, RAI, and ERP are presented in Table 2.

 

TABLE 2. Characteristics of all female patients with RRI, RAI, and ERP

Age and Intussusception

Analysis of variance revealed that the difference in the median age of the proctographic groups (RRI, RAI, and ERP) was statistically significant (p = 0.0001; Table 2) between the RRI group and the RAI/ERP groups, but not between the RAI and ERP groups. Intergroup comparison confirmed the statistical difference in age between RRI/RAI (p= 0.001, MWU) and RRI/ERP (p = 0.003, MWU) groups.

Symptoms and Intussusception Type

Rectorectal intussusception and RAI patients were symptomatic with fecal incontinence or obstructed defecation for a median period of 60 months, whereas ERP patients had symptoms for only 36 months (p = 0.006, ANOVA). The difference in the duration of symptoms, however, was only statistically different between the RRI and ERP groups (p = 0.0002, MWU).

Anorectal Physiological Parameters

Sphincter Complex Dysfunction.

The proportion of patients with defective internal and external anal sphincters (IAS and EAS) was statistically different (p < 0.0001, ANOVA; Table 2) between the proctographic groups with a higher percentage of ERP patients identified with defective IAS and EAS (41.0%) in comparison with the other 2 proctographic groups (RRI, 16.0% (p < 0.05, MWU); RAI, 22.6% (p < 0.05, MWU)). Further analyses assessing the presence of an isolated sphincter defect did not show any statistical difference between the proctographic groups when EAS defects were considered (p = 0.19, ANOVA), but was statistically significant for IAS defects (p = 0.018, ANOVA) in the comparison of the RRI and ERP groups (p= 0.017, MWU).

Median maximum resting pressure decreased in a linear fashion across the proctographic groups with the lowest pressure identified in patients with ERP (41 cmH₂O; range, 7–168) and the highest pressure identified in patients with RRI (77 cmH₂O; range, 5–240) (p < 0.001, ANOVA; Table 2). The difference in resting pressures was confirmed to be statistically significant between the RRI and RAI (p = 0.002, MWU) or RRI and ERP (p < 0.001, MWU) groups.

The median maximum squeeze pressure also appeared to vary across the proctographic groups (p < 0.0001, ANOVA) when comparing RRI (56 cmH₂O; range, 0–378) and ERP (40 cmH₂O; range, 5–150; p < 0.0001, MWU) or RAI (50 cmH₂O; range, 8–221) and ERP (p < 0.011, MWU) patient groups.

Rectal Sensation/MTV.

The MTV also varied across the groups (p = 0.007, ANOVA) with the difference confined between RRI (180mL; range, 44–660) and RAI ((160mL; range, 60–420) (p = 0.004, MWU)). The percentage of hyposensate patients, however, did not differ (p = 0.19, ANOVA).

Pudendal Neuropathy.

No statistical difference was observed in the proportion of patients with unilateral or bilateral pudendal neuropathy across the proctographic groups (Table 2).

Uni-/Multivariate Analyses.

Multivariate models were developed by an iterative approach from information derived from univariate analyses with the use of stepwise omission of covariates to improve the models. Analyses demonstrated that the only determinant for each group was age, with an association between increased age and RI demonstrated, albeit with a poor correlation (95% CI, 0.958–0.990; p > 0.001 (model probability >x² = 0.01; r² = 0.006)).

Patients With Repeated Proctograms.

A total of 32 patients were identified to have undergone repeat proctography as part of their ongoing clinical care for persistent symptoms of fecal incontinence (n = 6), obstructed defecation (n = 4), or both (n = 22). The median interval between the initial and repeat proctographic study for all 32 patients was 2.5 years (range, 1–7 years). Twenty patients had RRI on the initial and repeat proctographic study. Six patients demonstrated progression from RRI to RAI (n = 5) or ERP (n = 1) within a median period of 2 years (range, 1–4 years). An additional 5 patients were identified to have RAI on the initial proctogram followed by RRI on the follow-up study, having undergone abdominal or perineal rectopexy in the interim. Another patient was diagnosed with ERP on follow-up proctography despite surgical intervention. Consequently, in this study, 5/26 (19.2%) patients with initial RRI progressed to RAI within 2 years, whereas the progression from RRI to ERP was 3.8% (1/26).

DISCUSSION

In this study, there was a weak association between age and RRI, RAI, and ERP. In addition, symptom duration did not differ between the proctographic groups with the exception of RRI patients (60 months; range, 1–936) and ERP patients (36 months; range, 2–732). Furthermore, the progression rate from RRI to ERP, estimated from patients who underwent serial proctograms, was only 3.8%, although this rate could potentially increase over a longer time period, because 19.2% of patients were identified to progress from RRI to RAI within 2 years. The progression rates described by our study are similar to those reported by other authors. For example, Mellgren et al¹² reported that 6.7% of patients operated on for RI developed ERP after a mean follow-up of almost 6 years. In comparison, Choi et al¹³ reported that only 3.8% of patients operated on for intussusception progressed to ERP, although the follow-up period for this patient group was not as long as that of Mellgren et al (4 years). Both of these studies, and our own data, as well, are characterized by short follow-up so that the true progression rates may be even higher. However, the effect of selection bias cannot be ignored, because the subjects in all of these studies are symptomatic patients referred to a pelvic floor clinic. In addition, in the studies by Mellgren and Choi, the more severely affected patients were recruited without randomization, potentially resulting in an artificially higher progression rate.

Previous studies have suggested that RI is a precursor to ERP, because younger patients often present with intussusception, whereas older patients present with overt rectal prolapse.¹¹ The association between age and prolapse type reported by Wijffels et al¹¹ involves a considerably smaller female cohort (n = 471). Furthermore, the correlation coefficient between prolapse type and age for nulliparous (Pearson coefficient = 0.17; p = 0.16) and parous females (Pearson coefficient = 0.41; p < 0.0001) in that study was also poor.

Consequently, the available evidence supporting the hypothesis that intussusception is a condition progressing from RRI to ERP, on the basis of age difference, is weak and inconclusive. A degree of progression, demonstrated by repeat sequential proctograms, appears to occur, but the likelihood of ERP is difficult to accurately predict for patients with RRI/RAI because sample sizes tend to be small and studies are characterized by selection bias. The progression rates reported in the published literature are low, but this may be because the end point of the condition (ERP) is reached infrequently and only when a combination of risk factors coexists for a sufficiently long period of time. Perhaps the development of ERP may primarily involve a different mechanistic pathway associated with the presence of a congenital deep pouch of Douglas, loose fixation of the rectum within the sacral hollow, or a congenital collagen abnormality.⁵,¹⁹ Therefore, long-term observational proctographic studies are required to obtain important mechanistic data and draw definitive conclusions.

Assessment of rectal sensation in this study by using MTV as a physiological end point revealed that the MTV was greater in RRI patients than in RAI patients. This finding may be due to a perceived decrease in compliance resulting from infolding of the rectal wall as a result of high-grade intussusception, with early intramural mechanoreceptor activation, and concomitant loss of volume within the rectal ampulla. However, the proportion of patients with documented rectal hyposensitivity was between 8% and 14% for all proctographic groups and was not found to be statistically different. This may suggest that blunting of afferent rectal sensory pathways may be unrelated to the development of RI/ERP and thus may not be a major contributing factor in the development of RI or ERP in any of these patient groups. Other authors report similar rates of rectal hyposensitivity in patients with high-grade intussusception, further supporting the above theory; thus, it may be reasonable to conclude at the present time that blunted rectal sensory function is not associated with the development of RI/ ERP per se, but is, in fact, more likely to result as a consequence of the risk factors that predispose patients to develop RI/ERP.²⁰

Manometric measurement of anal pressures demonstrated that patients with ERP tend to have lower resting pressures than patients with RRI. This observation may be explained by persistent activation of the anorectal reflex in those patients with ERP rather than as a result of injury to the anal sphincter complex, although a greater proportion of patients with ERP were found to have defective IAS and EAS in comparison with patients with RRI or RAI.

This study is, to our knowledge, the largest to date, and although it is limited by its retrospective nature, it has contributed to current evidence that ERP may develop by a natural progression from RI, but that this may occur rarely and only when a combination of risk factors coexist for a long period of time. However, nulliparous young women without any identifiable risk factors also develop ERP, suggesting that a second mechanistic pathway may provide an alternative route.³ Available data suggest that a possible congenital weakness of the pelvic floor musculature⁴,⁸ and/or associated nerve supply may be implicated. Consequently, the contribution of disorders relating to the deposition and organization of pelvic connective tissues will require investigation in keeping with a previous report demonstrating that an abnormal collagen ratio in nulliparous women may be associated with stress urine incontinence.²¹ Perhaps there are 2 types of patients who develop full-thickness rectal prolapse: 1) patients with severe congenital collagen abnormalities that develop prolapse at a younger age, and 2) patients with a mild/moderate collagen disorder who acquire risk factors contributing to the development of RI and subsequently develop external prolapse in older age. Nonetheless, it is clear that future studies will need to use a combination of radiological (eg, dynamic MRI), histological (eg, biopsy), and neurological (eg, nerve conduction studies) assessments to appreciate in more detail the pathophysiological and mechanistic steps involved in the development of intussusception/prolapse for all patient groups.

REFERENCES