Felicity J. Creamer and B. James Mander
The emergency presentation of anorectal pathology constitutes a significant proportion of the general surgeon's workload. The problems encountered range from the acute pain of thrombosed haemorrhoids and perianal sepsis to the management of massive bleeding, trauma and foreign bodies. The management of many of these conditions is frequently straightforward; however, the close proximity of pathology to the structures involved in continence does require a degree of care and reflection. It is therefore imperative to ensure that any surgical intervention is appropriate, timely and minimally disruptive.
Understanding the pathophysiology and rationale for treatment in anorectal disease is impossible without sound knowledge of the anatomy of the anal canal. This is a 3- to 4-cm-long tube running downwards and backwards from the anorectal angle to the anus. It is divided in half by the dentate line, above which the canal is lined with hindgut-derived columnar epithelium and innervated by the autonomic hypogastric nerves, sensitive only to stretch. The section of the canal below the dentate line is lined with stratified squamous epithelium that merges at the anus with the perianal skin, and its nerve supply is derived from the somatic inferior rectal nerve making it sensitive to pain, pressure and temperature.
The canal is surrounded by a funnel of muscle essential in maintaining faecal continence. The inner muscle layer is the involuntary internal sphincter muscle, derived from the circular muscle of the rectum. Beyond this is the voluntary external sphincter, formed from striated muscle, continuous at its superior edge with the levator plate. Between these two muscle layers is the intersphincteric space, which contains mucous-secreting anal glands. Ducts from these glands open into the anal valves at the dentate line.
Three submucosal anal cushions are found, usually at the 3, 7 and 11 o'clock positions within the anal canal and contain fibroelastic tissue and arteriovenous anastamoses. In health they appose to form a tight seal within the canal, which helps to maintain continence, but in some people these can enlarge to form troublesome symptomatic haemorrhoids.
Perianal abscesses are defined as any collection of pus in the perianal tissues and are very common, with operative treatment being required in about 1 in 5000 people in Scotland annually.1 They occur predominantly in adults, with a peak incidence in the third and fourth decades of life, men being affected two to three times more frequently than women. Patients present with signs and symptoms of acute inflammation, with pain being the most common symptom. In approximately one-third of patients they are associated with a persistent fistula-in-ano that, if left untreated, can result in recurrent infections.2
Primary perianal abscesses are most commonly caused by infection within the anal glands (the cryptoglandular theory) with the common causative organisms being gut-derived enterococci.3,4 The resulting suppuration can spread in several directions, resulting in infection in a variety of anatomical spaces (Fig. 11.1). The relative frequency with which abscesses occur in the various anatomical locations is shown in Table 11.1.5 Primary abscesses can also be caused by suppurating skin infections, including carbuncles, furuncles and infected apocrine glands. The responsible bacteria in these cases is almost invariably staphylococcus and as these abscesses do not communicate with the anal canal they are not associated with fistula formation.6 Studies have shown that taking a swab for culture at the time of abscess drainage can predict whether a fistula is likely to be present and thus guide decision-making regarding future management.6,7
Location of anorectal sepsis by anatomical site
Data from Ramunjam PS, Prasad MI, Abcarian H et al. Perianal abscesses and fistulae: a study of 1023 patients. Dis Colon Rectum 1984; 27:593–7. With kind permission from Wolters Kluwer Health.
FIGURE 11.1 The spread of anal gland infection and common sites of anorectal sepsis. Infection of the anal gland within the intersphincteric space can spread in a variety of directions (see left side of diagram), resulting in abscess in a number of classical sites: 1, supralevator; 2, intersphincteric; 3, submucosal; 4, perianal; 5, ischiorectal. Note also the possibility of circumferential extension of sepsis (‘horseshoeing’) in the intersphincteric, ischiorectal and supralevator planes.
It is recommended that a swab from the abscess cavity is sent for culture to determine the likelihood of a subsequent fistula.6,7
Secondary abscesses are much less common, accounting for just 10% of presentations.8 They are the manifestation of distinct underlying disease, with inflammatory bowel disease (Crohn's disease in particular), colorectal neoplasia, diabetes mellitus, AIDS and tuberculosis all being potential causes. They can also occur as a complication of haemorrhoid surgery or as a consequence of trauma from foreign bodies.
On examination the abscess can usually be seen as a red, tender, fluctuant swelling near the anal verge; however, ischiorectal abscesses often present as a less distinct brawny swelling on one side of the anus and intersphincteric abscesses cannot usually be seen externally at all. Although this last type of abscess can be felt through the anal wall as a smooth, tender collection, digital rectal examination is usually excruciatingly painful without anaesthesia. This diagnosis should therefore be suspected in patients with severe anal pain and fever. In a few patients (particularly those who are immune-compromised or with diabetes mellitus) the abscess can be associated with cellulitis, which can progress to life-threatening necrotising infection if not treated promptly.9
All patients presenting with anorectal sepsis or pain should have an examination of the anorectum, including proctosigmoidoscopy. It is our opinion that this can usually only be performed satisfactorily under general anaesthesia. As the diagnosis is usually obvious, few people would routinely recommend preoperative imaging (although some advocate its use, arguing that by identifying cavities and fistulas there is a reduction in the incidence of recurrence10). However, if the diagnosis is unclear (such as in intersphincteric or supralevator abscesses) or in cases of recurrent sepsis, imaging is very useful.
Magnetic resonance imaging (MRI) is an accurate method of identifying collections of pus and fistula tracts.11,12 It is commonly used in investigating complex fistula disease in Crohn's disease but is rarely needed in the acute situation. Endoanal ultrasound is another sensitive method of identifying collections of fluid close to the anal canal and can also be used to identify fistula tracts (Fig. 11.2). It remains a technique practised by only a few specialist surgeons and so is rarely available in the emergency situation. However, it has been shown that in experienced hands it is at least as accurate as MRI and its use is likely to increase.13,14
FIGURE 11.2 Endoanal ultrasound examination carried out on a patient under general anaesthesia. This patient presented with severe anal pain and perianal induration but without any specific area of fluctuation. The ultrasound demonstrates an extensive ischiorectal abscess cavity (c,d) extending around the anal canal (a, internal sphincter; b, external sphincter). With thanks to Mr Mike Hulme-Moir, previous Clinical Fellow in Colorectal Surgery, Royal Infirmary, Edinburgh.
Treatment must be aimed at draining the abscess, thereby removing the source of sepsis. This should be done whilst minimising damage to the sphincters, preventing recurrence and minimising hospital stay. In most patients this involves making a linear incision over the cavity, gently breaking down any loculi and then dressing appropriately. There is no role for treatment with antibiotics unless the patient is immune-compromised or there is evidence of florid cellulitis or suspicion of necrotising infection. Although in North America this drainage is commonly undertaken under local anaesthesia, the more thorough examination, drainage and treatment that can be performed under general anaesthesia is, in the opinion of the authors, preferable – and probably prevents a significant proportion of re-operations. A large retrospective case study of 500 patients treated for perianal abscess at the Mayo Clinic was reported in 2001 and revealed a 7.6% re-operation rate.15 The reasons for re-operation included incomplete drainage of the abscess cavity at the first operation, missed abscesses (most often posterior collections) and postoperative bleeding. There was no association reported between patient variables (such as age, immune suppression or diabetes) and so it must be concluded that surgical error was the only reason for these findings, emphasising the need for a thorough primary examination.
Historically, it was suggested that thoroughly ‘deroofing’ the abscess cavity with a wide cruciate incision was beneficial but this does little other than giving the patient a larger wound that will take longer to heal.16 Similarly, a short-lived enthusiasm for primary wound closure after incision has been abandoned by most surgeons as studies have shown that it offers little immediate benefit in terms of time to wound healing and probably increases the chance of recurrent sepsis.17
If the abscess cavity is very large, an alternative to making a huge incision is to insert a de Pezze or Malecott catheter via a smaller skin incision. This was looked at in one study, which showed that of the 91 patients who underwent this treatment compared to the 54 who underwent conventional treatment, hospital stay was shorter (1.4 vs. 4.5 days) and the need for community dressing shorter, with no disadvantages seen at long-term follow-up.18
The management of specific abscesses is shown diagrammatically in Fig. 11.3. Simple perianal abscesses should be drained and the cavity gently curetted (Fig. 11.3a). With ischiorectal abscesses, the cavity is often huge (Fig. 11.3b). The cavity should be incised as near to the anal verge as possible to bring the external opening of any subsequent fistula close to the anal verge and thus minimise the trauma of subsequent fistulotomy should it be required. A large horseshoe abscess in the ischiorectal space is better drained through multiple short incisions than one large circumferential incision. As mentioned earlier, the size of the drainage wound can be minimised if a drainage catheter is used. Intersphincteric abscesses require drainage into the anorectum, with excision of part of the internal sphincter (Fig. 11.3c). Submucosal abscesses, although rare, are drained into the anal canal. Supralevator or pelvic abscesses must be drained with care and ideally not through the perineum or a high fistula-in-ano will be created. In true pelvic abscesses unrelated to spread from the anal glands, drainage can be achieved into the rectum or vagina. If the abscess is related to pelvic pathology, the primary disease process will need to be excised along with drainage of the pus.
FIGURE 11.3 Management of specific abscesses. (a) Perianal abscess is treated by excision of a small disc of skin and curettage of the cavity. (b) Ischiorectal abscess may require excision of a substantial amount of tissue to facilitate drainage. The alternative is to introduce a drainage catheter through a small stab incision. (c) Intersphincteric abscess is treated by excision of the mucosa and internal sphincter overlying the abscess. Such an abscess should not be drained through the perineal skin or a high fistula will result.
After the abscess has been opened and drained it is common to ‘pack’ the resulting cavity with absorbent dressing material, which must then be removed and replaced by community nurses. However, there is no good clinical evidence that this practice is beneficial to the patient and indeed, in our opinion, serves only to cause discomfort and inconvenience.19
Perianal abscesses are best incised under general anaesthesia with the cavity left open. Packing the abscess cavity is clinically unnecessary and unpleasant for the patient.15,17,19
Studies have shown that perianal abscesses are associated with fistulas in about 60% of patients;7,20 however, only 29–37% of these persist after the acute inflammation has resolved.2,21,22 A persisting fistula tract increases the chance of a recurrent abscess and can cause perianal discharge, itch and pain even in the absence of an acute infection. However, identifying a fistula when acute inflammation is present can be tricky, especially for the less experienced surgeon. As noted previously, a swab from an abscess cavity that grows gut- rather than skin-derived bacteria would suggest a communication to the bowel.6,7 These patients can then be reviewed in the clinic after the acute sepsis has resolved and a second-look examination under anaesthesia (EUA) booked if indicated.
However, the cryptoglandular theory of anorectal sepsis would suggest that the presence of intersphincteric pus would definitively predict an underlying fistula tract. This was tested in a 1994 prospective study of 22 patients. A radial drainage incision was made and extended into the intersphincteric space. Careful examination was then made to determine whether there was a fistula tract. It was found that intersphincteric sepsis predicted a fistula with 100% sensitivity and specificity. Patients were followed up for 38 weeks with no adverse outcomes reported. This has the advantage over microbiological analysis of the abscess (which is very sensitive but just 80% specific) in that no second procedure is needed.23 However, it must be noted that this study was performed by two highly experienced colorectal surgeons and there is no evidence that similar results would be safely obtained by non-specialists. The probing and opening of tracts within friable, oedematous tissue, with the possibility of creating false passages, may do disproportionate damage to the sphincters and thus continence.
The majority of anorectal abscesses are adequately treated with incision and drainage alone and if a fistula tract is not obvious it should not be sought.20,21
So should synchronous fistulotomy ever be performed? A recent Cochrane review of six randomised control studies from five different centres from 1987 to 2003 looked at concomitant fistula surgery at the time of abscess drainage in terms of recurrence, need for further surgery and postoperative incontinence.24 However, as would be expected, the eligibility criteria and treatments offered varied considerably between the trials, making definitive recommendations difficult. Three studies also report fistula rates of 83–90%,20,25,26 which is higher than would be expected, raising questions about whether iatrogenic tracks may have been created.
The majority of the studies included in the review dealt only with the surgical treatment of low fistulas. Of these, one randomised patients after the initial abscess drainage, with fistulotomy performed as a second procedure on day 3 of the acute admission.25 Most studies excluded those with recurrent anorectal sepsis, previous surgery and inflammatory bowel disease (IBD).24
All the studies showed that recurrence was less likely after fistula surgery (risk ratio 0.07–0.24), although follow-up times varied. Only two studies looked at short-term incontinence, with one (which included only low fistulas, n = 52) reporting no clinical incontinence in either group despite anal manometry revealing a transient reduction in the anal resting pressure in the fistulotomy group (76.3 mmHg vs. 91.1 mmHg), which had disappeared by week 12. Squeeze pressures were unaffected.26 The other group (who had included high trans-sphincteric and suprasphincteric fistulas, which had been treated with cutting setons) reported transient incontinence of flatus in 3/100 of the control and 15/100 of the intervention group. For those who underwent drainage alone this had entirely resolved by 6 months; however, in the fistula surgery group, 6% were still having problems after 1 year.20
Long-term continence rates were documented in five studies and were normal in four of them following simple drainage alone at 1 year or later. Of the studies performing low fistulotomy, incontinence of flatus was 0/24,27 8/2025 and 0/24.26 The study that included high fistulas reported 2/100 incidence of incontinence for flatus and 4/100 for liquid incontinence with urgency.20 The final study had performed fistulectomy and partial internal sphincterectomy, and they reported flatus/liquid incontinence in 6/32 of the control and 13/34 of the intervention groups, with four and one patients from each group, respectively, lost to follow-up.28 All these data combine to give a risk ratio of 2.64 for long-term incontinence following synchronous fistulotomy.24
The overall conclusions of the Cochrane meta-analysis were that synchronous fistulotomy is appropriate for low, uncomplicated fistula tracts. It should not be performed for high fistulas or anterior fistulas in women. It should also be avoided for groups where the risk of incontinence is high (e.g. those who have undergone previous anorectal surgery or have IBD).24
Synchronous fistulotomy can be performed with care for low, uncomplicated fistula tracts to reduce the risk of abscess recurrence and further operations.24
Management Of Secondary Perianal Sepsis
These abscesses should be drained like any other but definitive treatment will require resection of the malignant lesion. However, it is likely that the presence of anorectal sepsis indicates tumour spread outwith the bowel wall, which has implications for further treatment. Significant malignant fistulas will require a defunctioning stoma as other treatment will not be successful.
Inflammatory bowel disease
All abscesses in patients with IBD should be drained and seton sutures inserted into any fistulous tracts as necessary. MRI is often helpful in demonstrating fistula tracts. Recurrent disease is often very hard to treat and an expert opinion should be sought.
This life-threatening condition requires urgent wide debridement of all involved tissue and high-dose antibiotics. A defunctioning stoma may have to be formed to allow healing and several repeat examinations under anaesthetic are often necessary to ensure that only healthy tissue remains. Microbiology advice should be sought for antibiotic choice.
Anorectal Sepsis In Children
Anorectal sepsis in children is uncommon. One of the largest series in the literature is from Edinburgh, which reported 69 patients in a catchment population of 1 million over a 10-year period.29 The median age for development of the abscess was 3 years (range 1 month to 12 years), with the male to female ratio 9:1. In total, 24 patients (38%) presented with recurrent sepsis after simple incision and drainage, but in only half of these was a fistula found. The study was unable to relate pus culture to the presence of a fistula. One study has shown that non-operative treatment with needle aspiration and antibiotics was satisfactory in the management of 36 children over 2 years of age, with only four requiring operative incision.30However, a recent large study from New Zealand has shown that the risk of recurrence is significantly reduced by incision, drainage and fistulotomy where possible.31
A pilonidal sinus arises from infection within a hair follicle, usually in the natal cleft. It is an acquired condition caused either by an ingrowing hair, which then sets up a local foreign body reaction, or by obstruction of a hair follicle, which then ruptures into the surrounding tissues. It affects twice as many men as women and is more common in the hirsute.32 In the Second World War it was known as ‘Jeep disease’ due to its prevalence amongst the drivers in the United States army. The usual emergency presentation of pilonidal disease is with an acutely inflamed tender swelling adjacent to the natal cleft. Midline pits are usually visible. The causative organism is, as with many skin infections, staphylococcus but mixed anaerobes are not infrequently cultured.33
In the emergency setting the objective of treatment is to drain the abscess, which usually relieves the acute symptoms and prevents spreading sepsis. Excision for the whole sinus tract at the same time as abscess drainage has been attempted but this has been associated with recurrence rates of up to 60%.33 In fact, primary drainage alone has a similar recurrence rate of around 50%,34 but with a much smaller initial wound.
A comparative study from Israel assessed 58 patients with acute pilonidal abscess; 29 patients were treated with incision and drainage, the other 29 with wide excision (without closure). The results are displayed in Table 11.2. The risk of recurrent pilonidal suppuration was similar between the two groups, although those who underwent excision had a longer time off work.34 It is therefore recommended that excision of the fistula tract is not undertaken synchronously.35 It is the authors' usual practice only to offer formal excision after the second presentation with a pilonidal abscess and this is always performed as a separate procedure after the acute sepsis has settled. To ensure optimum wound healing it is recommended that the incision is made away from the midline and, as with other abscesses, that packing the wound with dressings is avoided.
Results of treatment for acute pilonidal abscess
*Values expressed as median (range).
Data from Matter I, Kunin J, Schein M et al. Total excision versus non-resectional methods in the treatment of acute and chronic pilonidal disease. Br J Surg 1995; 82:752–3. © British Journal of Surgery Society Ltd. Reproduced with permission. Permission is granted by John Wiley & Sons Ltd on behalf of the BJSS Ltd.
Fifty per cent of all pilonidal abscesses will be cured with incision and drainage alone.34 Definitive surgery should be reserved for those who have recurrent disease and performed in the elective setting.33,35
Acute anal fissure
A primary anal fissure is a benign, superficial ulcer within the anal canal. Patients present with severe pain on defecation, which is caused by chemical irritation of the ulcer and spasm of the internal sphincter. Some patients will also complain of bright red bleeding caused by irritation of the ulcer bed. Although most patients with anal fissures will be seen in the outpatient clinic, some present acutely with an exacerbation of pain or just out of desperation with their chronic symptoms. It is often an easy diagnosis to make on inspection; gently parting the buttocks will usually reveal the lower edge of the fissure, most commonly in the posterior midline (6 o'clock) position.
The key to treatment is to stop the muscle spasm, which can be achieved by topical administration of smooth muscle relaxants such as glycerine trinitrate (GTN)36 or diltiazem,37 which reduce sphincter pressure and aid fissure healing. Diltiazem has a significantly lower side-effects profile than GTN and so should be used preferentially, if available.38 These treatments typically take weeks to exert their full effect, but one study has shown that they can be useful in the acute setting – most patients presenting with fissures are too sore to tolerate digital rectal examination but administration of sublingual GTN allowed 13 of the 16 patients to be examined.39 This allows more sinister pathology to be ruled out without recourse to admission and anaesthesia in some patients.
Medical treatment has excellent short-term results, leading to the healing of 90% of acute fissures without the need for surgery.40 However, the long-term outlook is less encouraging, with one meta-analysis showing recurrence rates of up to 50%.41 Lateral internal sphincterotomy (LIS) is much better at reducing recurrence (rates of about 2%) and although some studies have raised concerns about continence disturbance, it has been shown that a well-planned procedure in selected patients has very little effect on continence.42,43 This, however, has no place in the acute setting and should be reserved for patients with chronic problems. Botox (botulinium toxin) injections into the internal sphincter are something of a middle ground, affecting a temporary sphincterotomy to allow fissure healing. These have again been mostly used for patients with chronic problems but a recent Eygptian randomised control trial has shown good results when used for acute fissures.44
It is important not to underestimate the pain caused by anal fissures and if patients are in pain for long periods of the day or unable to sleep at night an EUA should be performed to exclude occult sepsis and botox and local anaesthetic infiltrated into the internal sphincter and under the fissure, respectively, to provide symptomatic relief.
Most anal fissures will heal with conservative topical treatment.36,40 However, an EUA should be performed for those with severe unremitting symptoms. Definitive surgery should be reserved for those with recurrent disease.41,42
Haemorrhoidal disease is very common, causing symptoms in approximately 5% of the population.45 Haemorrhoids can be divided into internal, which originate above the dentate line, and external, which originate below the dentate line and are thus covered in mucosa and sensitive to pain. Although haemorrhoids are associated with symptoms including perianal irritation and small amounts of bleeding, they do not usually cause pain and uncomplicated haemorrhoidal disease is rarely seen as an emergency. However, external haemorrhoids can thrombose spontaneously (this process may be associated with straining to defecate), and internal haemorrhoids may prolapse, strangulate and thrombose. This thrombosis, with associated oedema and sometimes necrosis of the overlying mucosa, combines to cause exquisite tenderness such that patients are often unable to sit, walk or defecate.
These tense, tender, purple swellings can be seen by simply parting the buttocks. Although immensely painful they tend to resolve spontaneously after 4–5 days and therefore many surgeons advocate non-surgical management, the mainstays of which are good analgesia, laxatives and topical treatments such as nifedipine.46 The main argument for this approach is that surgical treatment of haemorrhoids is likely to be painful for about the same amount of time as for natural healing, with little significant advantage in the longer term. However, a review by the American Society of Colon and Rectal Surgeons suggests that although those who present after 72 hours of symptoms are best served by non-operative treatment, those who present before this time would benefit from excision of the external component.45
However, definitive emergency surgical treatment can be difficult as swollen, congested tissues distort normal anatomy and restrict views. Although many retrospective and case-controlled studies suggest that outcomes are favourable when emergency haemorrhoidectomy is performed by experienced surgeons,47,48 the only prospective randomised controlled study to investigate this showed that non-operative treatment was associated with shorter hospital stay and less sphincter damage.49 If surgery is to be undertaken, several studies have shown that in the acute, as in the elective setting, a stapled procedure is associated with shorter hospital stays, reduced pain and earlier return to work.50–52 One case–control study from Singapore compared 204 patients who underwent emergency haemorrhoidectomy with 500 who underwent an elective procedure during the same time period. They demonstrated no difference in any of the assessed end-points (haemorrhage, stricture, incontinence and portal pyaemia) between the two groups (see Table 11.3).53
Results from comparative study on emergency and elective haemorrhoidectomy
Numbers in parentheses are percentages. NS, non-significant.
Data from Eu KW, Seow Choen F, Goh HS. Comparison of emergency and elective haemorrhoidectomy. Br J Surg 1994; 81:308–10. © British Journal of Surgery Society Ltd. Reproduced with permission. Permission is granted by John Wiley & Sons Ltd on behalf of the BJSS Ltd.
Thrombosed haemorrhoids will resolve with conservative management but longer remission may be achieved with surgery. If surgery is to be undertaken a stapled haemorrhoidectomy is preferable and has no greater associated risks in the acute situation if undertaken by a suitably experienced surgeon.49,53
Although per rectal bleeding is a common reason for acute surgical referral, major haemorrhage from an anorectal source is very rare. A series of lower gastrointestinal bleeds in an elderly North American population published in 1979 demonstrated an anorectal cause for massive blood loss in just four out of 98 patients.54 Bleeding generally comes from a colonic source but rectal cancer, haemorrhoids, proctitis, rectal varices, anal fissures and solitary rectal ulcer syndrome have all been implicated in massive lower gastrointestinal bleeds. The increased use of nicorandil for the treatment of ischaemic heart disease has led to an increased awareness of the rectal ulcers that this drug can cause as a side-effect and a recent case report has been published of a life-threatening bleed secondary to this.55
Worldwide the most common cause of anorectal trauma is childbirth, with 0.4% of all vaginal births complicated by a third-degree (into the external sphincter) or fourth-degree (into the rectal wall) tear.56 One prospective study using endoanal ultrasound to evaluate post-childbirth sphincter function has suggested that as many as 35% of women demonstrate damaged external or internal sphincters following vaginal delivery.57 These tears are often repaired in the labour suite with interrupted sutures to approximate the sphincters but results from this are poor, with significant levels of faecal urgency and incontinence persisting (up to 50% in one study).58 In an attempt to improve outcome, a recent randomised controlled study from Norway of 119 women with third- and fourth-degree tears compared this end-to-end approximation with an overlap technique for sphincter repair.59 Unfortunately, they found no significant difference between the two techniques for reported faecal incontinence at 12 months or on anal manometry.
Anorectal trauma can also occur as a result of penetrating injury, iatrogenic damage or secondary to foreign objects inserted into the anal canal. Injury sustained to the intraperitoneal rectum can sometimes be primarily repaired but if there is significant contamination, large injuries, devascularisation or nearby open fractures, resection and formation of a stoma should be preferred.60 Damage to the extraperitoneal rectum can also often be repaired primarily but proximal diversion may again be needed if the injuries are extensive.61 Sigmoidoscopy should always be performed if blood is seen in the rectal lumen or if an extraperitoneal haematoma is seen adjacent to the rectum at laparotomy.
Rectally inserted foreign objects and the innovative techniques used to remove them safely are extensively reported in anecdotal case reports in the world literature. These objects are most commonly inserted for sexual gratification and in most circumstances the patient has made an unsuccessful attempt to remove them before presentation. A review of these case reports suggests that in the majority of instances removal is possible under conscious sedation, either digitally for low objects or bimanually for those above the rectosigmoid junction.62When this fails, endoscopic extraction with or without fluoroscopic guidance is worth attempting. Some authors have reported success with various obstetric instruments and in one reported case of an irretrievable metallic ball, an electromagnet was employed.63 If all of these measures fail, or there is radiographic evidence of perforation, laparotomy is usually inevitable. In a series reported from San Diego, this was necessary in five of 64 patients presenting with impacted foreign bodies.64 It is recommended that all patients undergo sigmoidoscopy after extraction to ensure no damage to the rectal mucosa has been sustained.