Despite enormous interest in the development of an ideal operative procedure to treat anorectal malformations, the situation is as confusing as it was 30 years ago. The reasons could be inconsistency of the nature of the lesions found in patients along with multi-system associated malformations primarily affecting the prognosis of available procedures. High imperforate anus is a complex anomaly that requires a combination of careful preservation of structures and precise anatomic reconstruction for optimal results. A retrospective study, comprising a consecutive sample of female children with anorectal malformations treated over a period of 10 years, is presented. Of 130 patients, 83% (n=108) presented later than 3 months of age as they could pass stools through associated fistulae, of which the commonest type was found to be an ano- or recto-vestibular fistula (65%, n=83). Ninety-four cases (72%) had what are traditionally known as “low” anorectal malformations (perineal fistulae, anteriorly placed anus and anorecto-vestibular fistulae). Ten percent (n=14) had translevator, and 17% (n=22) were found to have “high” lesions. Seventeen percent of patients (n=22) had associated congenital malformations, predominantly in intermediate and high anorectal lesions, 40% of which pertained to the urogenital system. In total, 137 definitive operative procedures were done including Mollard's anterior perineal approach in 81 patients and posterior sagittal anorecto plasty (PSARP) in 38, enabling a comparison to be made of the functional results and complications in the two groups. Out of 24 post-operative complications noted in the series, 17 were found in the group who had Mollard's procedure carried out (including 3 deaths) as compared with 7 in these cases who had PSARP (two-sided P utilizing Fisher's exact test = 0.475).
Keywords: Anorectal malformations, ano-vestibular fistula, recto-vestibular fistula, Mollard, PSARP.
J.R.Coll.Surg.Edinb., 45, June 2000, 153-158
Management of imperforate anus has gone through evolutionary stages over centuries ever since Paulus Aegineta1 recommended the treatment of perineal incision followed by repeated dilatations in the 7th century AD. Despite unsatisfactory results and misery caused to these children because of strictures and closures, the method remained in vogue for almost 100 years until Benjamin Bell 2 changed this concept by attempting a precise dissection of the rectal pouch. In spite of advancement in knowledge and better techniques, the condition still carries considerable morbidity in the form of incontinence of stool after surgery due to improper placement of rectum with reference to the controlling muscles of the perineum.
In females, Mollard's anterior perineal approach3 has been used for over 20 years with various modifications, the latest being the anterior sagittal approach4,5 providing excellent exposure to the related structures.
The posterior sagittal anorecto plasty (PSARP), devised by Pena and deVries (1982), has revolutionized the management of anorectal malformations by providing complete exposure of the anatomy of the anorectal region during surgery.6 In addition, the method provides exact visualization of the fistula between the rectum and the female genital tract (rectovestibular, recto-vaginal) in addition to placement of the rectum within the confines of the striated muscle complex giving the best chance to achieve continence.7
The purpose of this article is to present comparative experience of anterior perineal approach and PSARP in 130 female children operated for various types and degrees of anorectal malformations, in the department of paediatric surgery at the Military Hospital Rawalpindi, over a period of 10 years (1988-1998).
All the consecutive female children (n=130) presenting for the management of anorectal malformations were included in the series. Patients who had diverting colostomies done elsewhere or had definitive surgery done in some other hospital, resulting in unacceptable results and needing revisional surgery (n=6), were also included. All the patients undergoing primary definitive surgery were examined, the fistulae probed, perineally scratched and invertogramed, where indicated. A radiograph of the pelvis was done in cases of doubtful sacral development. A routine fistulogram was not done and a distal colostogram only done in cases who presented without an obvious fistula (n=6). Patients undergoing definitive surgery were classified according to the types described by Pena8 (Table 1). Cases with a perineal fistula and anteriorly placed anus were considered together in one group, as were patients with anovestibular and rectovestibular fistulae as differentiation was not clear in more than half the patients. From 1988 to 1995, intermediate and low lesions (anteriorly placed anus, perineal and vestibular fistula) were treated by anterior sagittal modification of the Mollard approach (n=81), whereas after 1995, all such cases were treated by PSARP (n=38). Except in cases undergoing perineal anoplasty as a neonate (n=3) and a cutback procedure (n= 6), a defunctioning colostomy was done as a routine, and definitive surgery delayed until the age of one year. Perioperative antibiotics were used in all the cases using a combination of Gentamycin, Ampicillin and Metronidazole in appropriate dosage. Thorough gut lavage was an important pre-operative procedure. Functional results of surgery were documented and complications were noted.
Of 130 cases of anorectal malformations treated, inability to pass meconium and partial intestinal obstruction was the main cause for 17% (n=22) of the children being seen within 3 months of age, the youngest patient being one day old. Eighty three percent (n=108) presented later than 3 months because of the presence of fistulae through which they could adequately pass stools. Change-over to solid food and constipation (due to the fistula track being outside the sphincter mechanism) were the main causes for their presentation between 3 months and one year of age (n=70). The rest of the patients (n=38) either presented late because of social reasons or were admitted for re-do corrective surgery (n=38) due to unsatisfactory results after primary surgery.
The types of fistulae and lesions with which our patients presented is depicted in Table I. The majority (n=83, 65%) had associated ano- or recto-vestibular fistula, two patients did not fit into any type and were considered in the "unclassified" group. One of these had a ball of colon which opened into the dome of the urinary bladder, and the other had cloacal exstrophy. Six patients did not have an obvious fistula. Two of these had superficial lesions corrected by perineal anoplasty, one child presented with rectal atresia and two with a urogenital sinus defect, in addition to imperforate anus.
Table 1: Types of malformations encountered
| Classification | Number of cases |
|---|---|
| Anteriorly placed anus / perineal fistulae | 16 |
| Anovestiblar / rectovestibular fistulae | 83 |
| Rectovaginal fistulae | 20 |
| Cloacal malformation | 1 |
| Unclassified | 2 |
| Urogenital sinus | 2 |
| No fistula | 6 |
An attempt was made to establish the level of caudal arrest of the rectum combining physical examination, probing of fistulae and operative findings (Table 2). In our experience, external appearance of the perineum did not effectively correlate to the level of malformation in most of the children. Thirty-five associated malformations and lesions (Table 3) were seen in 17% of cases (n=22). Interestingly, the associated problems predominantly prevailed in intermediate and high anorectal malformations. Only 10% of these were seen in infra-levator low malformations. It is of further interest that 40% (n=15) of these associated malformations affected the urogenital systems and had a high percentage (60%) of vaginal defects.
Table 2: Level of malformations encountered
|
Level of the lesion |
Number of cases | Percentage |
|---|---|---|
| Infra levator | 94 | 72% |
| Trans levator | 14 | 11% |
| Supra levator | 19 | 14% |
| Very high | 3 | 3% |
Table 3: Associated malformations encountered
| Associated malformations | Number of cases |
|---|---|
| Short colon | 3 |
| Ball of colon | 1 |
| Malrotation | 1 |
| Rectal atresia | 1 |
| Urogenital sinus | 3 |
| Cloacal exstrophy | 1 |
| Exomphalos | 2 |
| Vaginal atresia | 4 |
| Septate vagina | 4 |
| Mongolism | 1 |
| Fallot's tetrology | 1 |
| Cardiac septal defects | 2 |
| Talipes equinovarus | 3 |
| Polydactyly | 2 |
| Absent uterus | 1 |
| Hypoplasia of toes | 1 |
| Ectopia vesicae | 1 |
| Hemivertebrae | 2 |
| Ambiguous genitalia | 1 |
A total of 137 operations were done as definitive procedures (Table 4) in the management of these cases. Preliminary colostomies, their closure and minor procedures (e.g like trimming of mucosa etc) have not been included, being adjuvant procedures which are part of the management. From 1988 to 1995, Mollard's anterior perineal operation was performed as a standard procedure (n=81) out of which 11% (n=9) needed an abdomino-perineal extension for their completion due to limitation of access, problems of mobilization, nature of the lesion and associated problems. From 1995 to 1998, all the patients were primarily operated by PSARP (n=38) out of which only two patients (5%) needed abdominal exploration for their completion. Three superficial cases were treated by perineal anoplasty whereas 6 patients with stenosed perineal fistulae (anteriorly placed) needed cutback procedures. These low lesions were managed without a protective colostomy.
Table 4: Operative procedures carried out
|
Operative procedures |
Number of cases |
|---|---|
| Mollard's procedure (anterior perineal) | 72 |
| Mollard's with abdominal extension | 9 |
| PSARP | 36 |
| PSARP with abdominal extension | 2 |
| Anal cut-back | 6 |
| Perineal anoplasty | 3 |
|
Laparotomy |
2 |
| Vaginoplasty | 4 |
| Vaginal pull through | 3 |
Patients who had a septate vagina (n=4) underwent vaginoplasty at the time of definitive anorectal repair, whilst cases of vaginal atresia were treated concomitantly at the time of PSARP or the Mollard procedure extending to abdomino-perineal exposure, where needed. Laparotomy alone was done in two cases. In one child, a ball of colon was separated from the bladder and opened on the abdomen as a colostomy. In the second, who developed shrinkage of the pulled down colon (probably due to relative ischaemia after PSARP), required a long-term sigmoid colostomy.
A total of 24 postoperative complications were noted in 130 patients (Table 5). Cases undergoing PSARP showed greater incidence of superficial wound dehiscence as post-operative sitting position caused direct pressure over the wound. Cases undergoing Mollard's procedure showed greater incidence of anal stenosis. Fibrosis and shrinkage of bowel occurred in one patient after attempted mobilization of the rectum while performing PSARP. Mild to moderate constipation was seen in 10% (n=13) of our patients.
Table 5: Comparative complications of Mollard's and PSARP operations
| Complications | Mollard's (n=81) | PSARP (n=38) | Total |
|---|---|---|---|
| Superficial wound dehiscence | 1 | 3 | 4 |
| Mucosal prolapse | 3 | 2 | 5 |
| Anal stenosis | 3 | 1 | 4 |
| Short-term incontinence (stools) | 2 | - | 2 |
| Short term incontinence (urine) | 2 | - | 2 |
| Incontinence of stools | 3 | - | 3 |
| Shrinkage of pulled down colon | - | 1 | 1 |
| Mortality | 3 | - | 3 |
| Total | 17 | 7 | 24 |
Six patients underwent re-do explorations for complications of primary surgery. One patient had a complete disruption of the perineum after an anterior perineal repair done elsewhere, and needed skin rotational flaps, in addition to PSARP correction. Four patients had to be revised by PSARP for incontinence with success being obtained in three patients. The sixth developed a non-yielding anal stenosis after Mollard's repair and was revised successfully by PSARP.
Mortality in our series was 2.3% (n=3) and was confined to the group undergoing Mollard's operation needing abdominal extension. It was directly related to complex associated malformations. In our series, patients subjected to Mollard's procedure had 74% incidence of complications (n=17), as compared with 26% (n=7) noted after PSARP (Fisher's exact test = 0.475, p=0.475).
After anorectal surgery, the functional results and evaluation of incontinence to urine and stool, if present, can only be done around 3 years of age.9 Evaluation of anorectal function is complicated and tests like anorectal manometry, anal resting tone estimation, anal squeeze pressure, rectal sensation and balloon distension are difficult to do in an un-cooperative child. It should also be remembered that some cases of soiling could be secondary to constipation rather than anal sphincter dysfunction. Our evaluation of patients, therefore, was limited to a thorough clinical investigation and detailed parental questionnaires and bowel-habit charts.
We reviewed all our patients every 3 months post -operatively, in order to diagnose early stricture formation or stenosis. The patients were later evaluated regarding functional results and continence according to categories laid down by Pena.9 Ninety five percent of our patients (n=123) have had normal continence, as far as feeling of urge to go and bowel action are concerned, up to the time of writing of this report. Seven patients had grade-I to grade-II soiling, which was temporary, settling within two months after operation. Three patients belonging to the Mollard's anterior perineal group had soiling of grade-III intensity in which re-exploration using posterior sagittal approach, revealed the rectum to be malpositioned with reference to the striated muscle complex and the controlling muscles. In two patients, the rectum had been pushed too far back without a significant anorectal angle, and in the third, it had been placed to the left of the muscle complex missing it altogether. Re-alignment with the help of a muscle stimulator significantly improved continence in these patients, two improving to grade-1 and the third to grade-II soiling.
Most of the children undergoing operative treatment passed formed stools on a daily basis. In 20% (n=27) of children the parents were concerned because they passed stools on alternate days only. Only 10% (n=14) of children had substantial constipation needing treatment as no other cause could be establish for the continued symptoms. Some of the patients who had PSARP are still under review.
Anorectal malformations in females have been treated in many different ways with variable modifications10,11 because of the realization that functional results of surgery were more important than the correction of the anatomical abnormality. Stephens (1986) laid maximum stress on the puborectalis sling being responsible for maintaining continence.12 Most of the perineal operations (e.g. Mollard's anterior perineal approach3 and their modifications4,5) were based on this approach. Rehbein (1959), on the other hand, was very much concerned with operative damage to the parasympathetic para-rectal nerves and brought forward the concept of endorectal pull-through which was rarely needed when dealing with female children (8%, n=10 in this series).13 Both Stephen and Rehbein did not give enough consideration to the external sphincter component, which was deliberately preserved by Yokoyama et al (1985), in addition to utilizing fibres of the internal sphincter.14
It has been consistently shown that low anorectal malformations predominate in female children5,15 (perineal fistula, anteriorly placed anus, and ano-vestibular fistula. [70% in this series, n=91]), as compared with high lesions. Mollard et al (1978) was the first to exploit this fact and devised the anterior perineal approach with the advantages of preserving the specialized anoderm of the anal canal and avoiding extensive pelvic dissection.16 The procedure also provided clear visualization and precise identification of the puborectalis sling with the advantage of obviating the need to change the position of the patient during surgery in cases where the abdomen had to be opened. This operation was easier and safer than the sacral approach and soon had many surgeons using the technique with variable modifications.4,17 Even Mollard (1989) himself modified his own technique by utilizing a tube of skin connecting the anoderm with the rectal pouch.18 This modification, however,was not popular.
Pena (1985) changed the concept of perineal musculature and postulated that instead of different muscle slings (i.e. puborectalis), the perineal muscles were continuous and the levators were joined to the external sphincter fibres with the help of the striated muscle complex.7 He further stated that all these muscles could be opened up in a sagittal plane without damaging their nerve supply. The procedure (PSARP) provided an excellent exposure to the perineal structures, associated fistulae could be handled under vision, the mobilized rectum could be placed precisely in the centre of the controlling muscles with the help of a muscle stimulator, and tapering of the rectum could be done, if needed. The procedure was based on sound scientific principles and was soon utilized for cases of incontinence resulting from failed primary repairs, and was associated with successful results.19,20
Recently, following the principles of PSARP, some centres have employed an anterior sagittal transanorectal approach (ASTRA)21 for clitoro-vaginoplasty in severely masculanized females and an anterior sagittal anorectoplasty (ASARP)22,as a re-do operation, in cases of failure of primary repair. The long-term results of both, as yet, are not available.
Vanden Baviere et al (1983)23 compared the results of Mollard's anterior approach 16 with the posterior trans-sacral approach, as devised by Stephens (1986)12 and found the anterior approach to have "undeniable" comparative advantage. Mulder et al (1995), on the other hand, compared functional results of PSARP with those of the Kiesewetter-Rehbein (1959) type of abdomino-perineal pull-through operation13 and concluded that for high and intermediate anorectal malformations, PSARP did not give better functional results than the pull-through operation.24 In their opinion, the prognosis was determined by other factors than the type of operation, notably the presence or absence of sacral defects. Our group, having done an equal number of both procedures (PSARP and abdomino-perineal pull-through) are of the opinion that PSARP is a much more precise procedure and saves considerable operating time, as compared with the pull-through procedure. Even if abdominal extension is needed in some cases, PSARP helps in predefining the exact pull-through route and precise rectal placement.
Heinen (1997) has reported a very high incidence of postoperative functional constipation, amounting to 47% in patients with an anteriorly placed anus and 50% in cases of vestibular fistula.25 Yeung and Kiely (1991) reported intractable constipation in 28% of their patients and highlighted the fact that in females, the lower the malformation, the higher was the incidence of severe constipation leading to over-flow pseudo-incontinence.26 The post-operative constipation rate in our patients was only 10% and in most of the patients it was of Grade I type, easily managed with oral medication. This discrepancy is attributed to different dietary factors in this country and a general increased incidence of loose stools because of lack of education regarding hygiene.
Contrary to previous teachings of operating on these children around the age of one year, a view has arisen suggesting that such malformations should be corrected much earlier. Pena (1988) has reported operating on these children around 2-3 months of age.8 Others have recommended such corrective procedures soon after birth hoping to achieve optimal results with lesser complications.4,27 Safer anaesthesia techniques, better monitoring during surgery and technical expertise have probably lead to a change in the older concepts. Dealing with a relatively un-educated population, the authors are still adhering to the standard age of one-year for all anorectal malformation corrections.
Hassink et al (1994) studied long-term results and quality of life in these patients (beyond 18 years-of-age) and found that 12% of patients remained socially restricted and handicapped and 24% could never have a lasting relationship because of such a handicap, thus, needing continued psychosocial support.28 Our follow-up has not been so long and we have not been able to establish such a relationship in our patients.
It is a well-known fact that the repair, however expertly carried out, cannot withstand the stresses of childbirth when these children grow up, become pregnant, and are allowed to go through normal labour. There have been instances (personal communications) where 3rd degree perineal tears have occurred in such individuals during the 2nd stage of delivery. In most of the instances, the obstetricians were not even aware if such a repair had been done in early childhood. Such a risk is commoner in developing countries with poor record keeping and lack of education. The authors regard delivery by caesarean section as mandatory in such females.
Mandatory use of muscle stimulation during PSARP has revealed that well-defined muscle structures were present in the anorectal region in all the patients studied and treated except those who had high anomalies in which they were comparatively weaker. PSARP allowed a satisfactory control of the procedure during surgery making it easier to avoid serious complications and mal-placement of the rectum, which in turn lead to superior anatomical and functional results. The extended exposure during PSARP also provided the best opportunity to create an anorectal angle, which is one of the important factors while considering continence. This important step was not possible to achieve or accomplish during Mollard's anterior perineal approach.
In all the cases with a fistula, the rectum normally deviates anteriorly to open into the perineum, vestibule or vagina. This results in the sphincter muscle fibres remaining un-used and probably accounts for the high incidence of lack of control on defaecation through this fistula before surgery. PSARP allows a precise splitting of this muscle complex to locate the mobilized rectum in its middle. This exact placement could not be properly accomplished using Mollard's anterior perineal approach.
Excessive mobilization of the rectal pouch with resultant excision of the terminal end can result unless the condensed circular muscle fibres constituting the internal sphincter.29 In order to take advantage of this structure, it is recommended that excessive tapering and trimming of the distal rectum should be avoided as it may lead to a low incidence of rectoanal inhibition reflex responsible for reduction in the anal resting tone and a high incidence of soiling problems.30 A more conservative mobilization and dissection results in the preservation of the distal rectum and less damage to its extrinsic nerve fibres. Anoplasty in these cases should also be made under slight tension to create a natal cleft and to achieve a slightly inverted anal canal which, otherwise, remains flat looking and prone to mucosal prolapse and soiling.
Copyright date: 18 May 2000
Correspondence: Mr M. A. Hashmi, 117 / 4 - C Race Course Road, Rawalpindi, Pakistan
E mail: hashmis@isb.pol.com.pk
©2000 The Royal College of Surgeons of Edinburgh, J.R.Coll.Surg.Edinb.