Anorectal Abscess Fistula

Surg Clin N Am 82 (2002) 1139–1151

Anorectal abscess fistula: what do we know? Richard Nelson, MD* University of Illinois at Chicago, 1740 West Taylor, m/c 957, Chicago, IL 60612, USA

It is possible that no condition has been written about more by surgeons over the expanse of time and world geography than anorectal abscess and fistula [1–4]. This broad concern is a testament both to the prevalent nature of these disorders and the difficulty they present to the surgeon in basic decision making. The reasons for these difficult decisions are obvious. It is necessary to resolve sepsis and symptoms associated with it, principally pain, and yet the procedures that one employs must also preserve function in the anal canal; that is, prevent incontinence.

Demographics Publications regarding anorectal abscess arise almost exclusively from single institution experience, which in some cases is very large [5,6]. These provide no population-based data, however. There is no denominator, so incidence cannot be calculated. Also, so many abscesses either rupture spontaneously and may not recur, or are lanced in a physicians office, that they never work their way into governmental records. Patients with fistula, however, at least in adulthood, are very unlikely to have spontaneous resolution. The vast majority of cases come to a surgeon’s attention. A trip to the operating room is this almost certainly recorded. In the city of Helsinki, Finland, the incidence of fistula, as determined by the number of people being treated for it, was recorded from 1969 to 1978: it was 8.6 cases per 100,000 people per year [7]. That would expand to an annual incidence in the United States of 20,000 to 25,000 cases per year being treated. This number conforms to data from the Ambulatory Care Survey of the National Center of Health

* University of Illinois at Chicago, 1740 West Taylor, m/c 957, Chicago, IL 60612. E-mail address: [email protected] 0039-6109/02/$ - see front matter Ó 2002, Elsevier Science (USA). All rights reserved. PII: S 0 0 3 9 - 6 1 0 9 ( 0 2 ) 0 0 0 6 3 - 4

1140

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

Statistics in the United States in 1979, when 24,000 individuals had anal fistula recorded as their primary diagnosis in American hospitals [8]. The number of recorded procedures has fallen precipitously since then. That is almost certainly due to the hospital discharge survey and the subsequent Healthcare Cost and Utilization Project (HCUP) recording only inpatient procedures. In 1999, only 3800 procedures for fistula were recorded in the United States by the HCUP [9]. Other aspects of fistula demographics included a gender ratio of roughly 2:1 men to women, both in the Helsinki and United States data [8,9]. This ratio is similar to rates in other large series [5,6]. This gender ratio is not maintained when one looks at other common conditions of the anal canal, such as hemorrhoids and anal fissure. It is also possible to estimate the incidence of perianal abscess from the fistula data. Several series have been published in which the rate at which a fistula was found at the time of abscess drainage, or the percentage of patients who had abscesses drained and were followed prospectively for subsequent fistula development (or persistence), was presented [6,10,11]. In these studies the incidence of fistula has varied from 26% to 37%. Working backwards, this would translate to an annual estimated incidence of anal abscess of between 68,000 and 96,000 cases in the United States. Though there are anecdotal observations concerning racial and socioeconomic prevalence of this condition, no reliable data exist that would allow calculation of a relative risk for various races or economic classes.

Etiology Fistula in the vast majority of cases arises from pre-existing abscess. Other etiologies of anal fistulae include superficial fistulae associated with anal fissure and anorectal trauma. The relative frequency of fistula is described in patients with ulcerative colitis and Crohn’s disease [12], though most surgeons now consider the presence of a fistula indicates that a patient with inflammatory bowel disease does have Crohn’s disease. A common cause of fistula described by Goligher that is now apparently vanishingly rare is tuberculosis of the anal canal [12]. The etiology of anorectal abscess is perhaps a bit more controversial. The anatomy of most fistulas with an internal opening at the anal gland in the crypt of Morgagni suggests that these abscesses originate with infections of the gland [6]. British writers felt that such an infection should result in a collection of pus in the intersphincteric space, between the internal and external sphincter. It is this belief that lead to a wide internal sphincterectomy being an integral part of the treatment of abscess and fistula in Great Britain, following the notion that drainage of this space was necessary to resolve sepsis adequately. Goligher, in a series in which he routinely performed sphincterectomy to explore the intersphincteric space, found very few patients with collection of pus in this region. He felt these were

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1141

convincing data against the cryptoglandular origin or anorectal abscess [12]. Of the two theories of the origin of fistula, the one that has won the hearts and minds of most colon and rectal surgeons is the one related to fistula anatomy and cryptoglandular origin. Risk factors/prevention Despite the extensive description of this topic, no body habitus, bowel habit, diet, or toxic exposure has ever been described that increases the risk for fistula or the alteration of which might decrease risk for fistula. In fact, the analytic epidemiology of the sort that is seen with great frequency for cancer of the colon or breast or coronary artery disease has never been performed for anorectal abscess or fistula. Why this disease should occur more frequently in men and why it should occur most commonly between the ages of 20 and 45 is unknown. Morbidity of abscess and fistula Almost all patients with abscess present with anorectal pain. Systemic sepsis may accompany this pain. Very rarely patients may present with no localizing symptoms or signs of sepsis, and the pain is found arising from the anal canal. Patients with fistula may present with either pain or sepsis, but far more commonly they have drainage either of pus or blood from the perianal region, unrelated to defecation. Sepsis may in some cases of perianal abscess be overwhelming and life threatening. When such is the case, especially in a diabetic (or immunosuppressed patient) with gangrenous skin changes, this constitutes a surgical emergency. Repeated debridement in lieu of a colostomy is the accepted treatment. For patients with anal fistula, the risk of recurrent sepsis is always present. In patients with chronic anorectal disease, the risk of developing cancer has been noted [13]. One study found six cancers in pre-existing fistulae, which had been present for a mean of 13.8 years. In none of these six was there an intraluminal rectal tumor. In three of the six the cancer was unsuspected at the time of surgery and only revealed to the surgeon when the pathologist reported three days later [13]. Several case control and cohort studies have been published that allow a numerical calculation of anal cancer risk related to fistula. In each case these have shown a statistically significant association of fistula with subsequent cancer development [14]. In those cohort studies in which the carcinoma diagnosis quickly followed treatment for anal fistula, one might suppose that the cancer arose simultaneously with the fistula or may have even caused it. Anal cancer tends to be an aggressive tumor, however [13,15]. For those individuals in whom the tumor diagnosis was made years after the fistula diagnosis, it seems more likely that the cancer arose in a pre-existing fistula. The pathophysiology of this malignant change has been studied extensively in other organs in

1142

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

which chronic inflammation often predates the diagnosis of cancer (such as in ulcerative colitis [16]). For a benign condition such as fistula, eradication of the inflammatory focus can be accomplished with minor surgery and low morbidity. This is a strong argument in favor of expeditious surgical treatment of patients with anorectal abscess and fistula.

Diagnosis The principal symptom of abscess is pain. The diagnosis of abscess is therefore established by first considering the differential diagnosis of anal pain. This includes, with abscess and fistula, anal fissure, hemorrhoidal thrombosis, proctitis, levator spasm, venereal diseases, and cancer. Because of the great distress that anal pain causes patients and the morbidity involved in delayed diagnosis of abscess, it is important to quickly determine the source of the pain. In many patients with abscess, examination will reveal the classic triad of rubor, tumor, and calor; that is, a hot red tender area adjacent to the anus. Many abscesses arise deep within pelvic tissue planes, however, and may not be clinically apparent on external examination. Also, patients with anal pain are resistant to digital examination of the anal canal, much less endoscopy, as they have tenderness as well as pain. Though there are many techniques that may infer the presence of an abscess, such as computerized tomography, magnetic resonance imaging, or radionucleide white-cell scanning, none of these are comparable in accuracy or therapeutic potential to an expeditious trip to the operating room for examination of the anal canal under anesthesia. This is best done under a regional anesthetic, allowing the patient to control his own airway, and in the prone jackknife position. There is no evidence that the performance of a spinal anesthetic in a patient with anorectal sepsis increases the risk of meningitis. If an abscess is suspected, especially if the patient is very septic, there is no reason to delay surgical exploration in order to explore alternative diagnostic techniques, or to attempt further resuscitation for resolution of the sepsis. For a diabetic with a gangrenous infection, early exploration, drainage, and debridement is the first step in controlling both his sepsis and diabetes. A special case for which clearly right answers seem elusive is a patient with acute leukemia and pancytopenia, perhaps either being prepared for or having just received a bone marrow transplant. Induration in the perineum is frequent in these cases and incision can be risky. Sometimes the anatomy of a fistula seems impossible to delineate at surgery. Insertion of a small catheter, such as those used for cholangiography, into the fistula with peroxide injection during surgery, or radiographs obtained during radio-opaque dye injection in radiology, can aid in the identification of an internal fistula opening [17]. Radiolabeled white-cell scanning has also been used in fistula patients, though its greatest usefulness is probably in ruling out proximal inflammatory bowel disease (Fig. 1) [18].

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1143

Fig. 1. (A) Anterior projection of a labeled white cell scan demonstrating sigmoiditis. (B) Posterior projection in the same patient showing the anorectal fistula.

Classification Classification systems for both abscess and fistula have been published. Each has its particular purpose. In the case of abscess, it is often unnerving to quiz vulnerable residents on the subtle differences between ischiorectal and intramuscular abscess. The purpose of classification is to alert the surgeon to the likelihood of occult components of the abscess. Even in cases where a superficial abscess is visible and drained, one must consider the

1144

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

possibility of a deeper collection of pus and be prepared for operative exploration to remove occult loculations if sepsis and pain do not rapidly resolve. Patients with perirectal abscesses incised and drained should be seen soon afterwards to assure that resolution is occurring. The classification of anal fistula has quite a different purpose aside from causing consternation to vulnerable residents. The anatomy of fistula is often complex, always interesting, and sometimes very surprising. Failure to appreciate this anatomy will likely result in recurrence or persistence of the fistula and the resulting morbidity. In addition, overly aggressive therapy for fistula can result in anorectal dysfunction; that is, incontinence. The classification systems for fistula therefore have been developed principally to assist in defining the anatomy of the fistula, as in the system used by Goodsall to describe the radial anatomy of fistulae, or to estimate the likelihood that incontinence might result from sphincter division in a lay open of a fistula, by gauging the vertical depth of sphincter encirclement by a fistula. The real essence of the Goodsall rule is that most fistulae originate in the posterior midline. Also, there is no perfect predictor of incontinence risk after any degree of sphincter division. A surgeon must be prepared to deal with a patient’s subsequent continence difficulties if he intends to treat a high anal fistula encircling a large percentage of the anal sphincter, particularly if that fistula is in the anterior midline in a woman.

Therapy At last we get to a portion of this discussion in which real evidence-based medicine can be considered. That is because only in this area are there a number of clinical trials, eighteen of which involve randomization of subjects in randomized controlled trials (RCTs) to various therapeutic groups. As such, this is by far the most important part of the discussion. Regarding abscess, there are a total of six RCTs that assess the feasibility of incision, curettage, and primary closure of perirectal abscess. Two of these trials involve abscess in all locations of the body, in which anorectal disease formed an important subset, but the results published did not separate those patients from the whole [19,20]. Two more studies applied curettage and closure to all patients and simply studied various antibiotic combinations [21– 23]. Two more studies are of greatest interest because they randomized patients to traditional incision and drainage or primary closure. The first of these was published over a generation ago when antibiotic choices were primitive. In fact, antibiotic therapy varied during the study as new drugs became available. It was a large study in which 109 patients were randomized to incision and drainage and 110 to suture with antibiotic cover. Patients with fistula were excluded. A major problem with this study was incompleteness of follow-up. Barely a third of the patients were seen up to a year after surgery. In this staggered follow-up, 26 patients with incision

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1145

and drainage had recurrence or fistula persistence, whereas only 8 of the patients with suture had recurrence [24]. Several years later a similar study was performed in which 41 patients were randomized to incision and 42 to closure with clindamycin cover. If a fistula was found at exploration, which it was in 5 of the first group and 3 in the second, a fistulotomy was done as part of the procedure. Though early results of this RCT suggested that recurrence or persistence of fistula was more common with conventional incision and drainage [25], a second publication presenting three-year follow-up turned the tables, showing almost twice as many patients recurred with prolonged follow-up after primary closure of the abscess [26]. A very primitive meta-analysis of this pair of publications shows an odds ratio favoring suturing of 1.62, but with a confidence interval of 0.88 to 3.09; that is, not statistically significant. This is not a reliable meta-analysis because of the very poor follow-up in the first study. A peculiar aspect of this practice of primary closure is the choice of suture material. This was often a heavy nylon, 0 or 2-0, with deep vertical mattresses, so that occlusion of the entire abscess cavity was accomplished. Though the authors described rapid resolution of pain, to have such large sutures in the perianal region and often embedded in the sphincter would seem very painful, not to mention the discomfort of suture removal. Nevertheless, it is curious that there were a number of authors that favored this approach. Yet in my career I have never seen it done, nor do I know of anybody in the United States that is doing primary suture of perianal abscess. The next therapeutic area of great interest is whether a definitive procedure should be performed for anal fistula at the time of abscess drainage. There are five randomized trials that address this issue, each rather different in its design, all published between 1987 and 1997. The first, by Hebjoern et al, randomized 18 patients to incision and drainage alone and 20 to definitive fistula therapy, which in his case was a fistulotomy [27]. Fistulotomy was in fact not done at the time of incision and drainage, but three days later, in the hope that some of the edema from the abscess would have subsided. A seton was used in some cases, and of the 18 patients randomized to incision, 2 had subsequent fistulotomy when recurrence became apparent. Recurrence developed in 3 patients who had incision and 2 who had fistulotomy. Incontinence of any degree was not seen in any of the patients having incision and drainage and in 8 of the patients with fistulotomy. It is concerning that of the 20 patients randomized to fistulotomy prior to surgery, 19 had it performed. The known prevalence of abscess in which fistula is found or develops suggests that this should have been done in only about 7 of the patients randomized to definitive fistula therapy. The next paper, by Schouten and van Vroonhoven [28], randomized 34 patients to incision and 36 to fistula therapy. The fistula therapy in this case was more radical: fistulectomy with partial internal sphincterectomy in order to drain the intersphincteric space. It is also of concern here that of the 36 patients randomized before surgery to have fistula therapy, all 36 had fistulas found at surgery

1146

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

and had this extensive fistula operation. Again, one would estimate that only about a third of them should have had fistulas found at the time of exploration. Recurrence developed in 13 of the incision patients, either as abscess or fistula, and only in 1 of the fistulectomy patients. Incontinence developed in 10% of the patients having incision alone and in about one third of the patients who had fistulectomy. Of those patients who developed recurrence after incision alone and had subsequent fistulectomy, 44% developed incontinence. On the other hand, it was found that 21% of patients presented a history of defecation disorder before surgery—almost ten times the population prevalence of incontinence in this age group [29]. Follow-up in this group ranged from 12 to 57 months. The study by Tang et al randomized 21 patients to incision and 24 to fistulotomy [30]. This is only study in which randomization only occurred after operative exploration and discovery of an internal opening, proving the presence of fistula. With one year minimum follow-up, 3 patients in the incision alone group developed recurrent abscess and none in the fistulotomy group did. Incontinence was also not seen in this study except in one of the incision patients. Ho and colleagues randomized 28 patients to incision and drainage and 24 to fistulotomy [31]. As in two of the three previous studies, randomization occurred before surgery, and yet 21 of the 24 patients randomized to fistula surgery ended up having a fistulotomy; that is, a fistula was discovered in 88% of this population with abscess. Postoperative fistula occurred in 7 of the 28 patients randomized to incision and drainage and one abscess occurred in this group. Only one abscess developed in the fistulotomy group. No incontinence was reported from either group with postoperative follow-up. Li and Yu randomized 32 patients to drainage and 39 to fistulotomy, though it is not clear when randomization occurred [32]. No incontinence was seen, and recurrence of fistula or abscess was found in 17 of the drainage group and in 1 of the fistulotomy group. A crude meta-analysis was performed of these five studies, calculating a stratified Mantel-Haenszel chi-square (EpiInfo 6.01; www.cdc.gov). For recurrence or persistence, an odds ratio of 13.2 was found, with a 95% confidence limit of 5.4 to 47, definitely favoring fistula surgery. On the other hand, the Mantel-Haenszel chi-square related to incontinence showed an odds ratio of 0.36 favoring incision and drainage ,with a confidence interval of 0.13 to 0.93. The reliability of these numbers is suspect for several reasons. The first is the large percentage of patients with abscesses in at least three of the five studies in whom fistulae were found when randomization occurred prior to surgery. The second is the variation in fistula therapy, particularly in the Schouten study. The third has to do with the nature of postoperative continence assessment, which will be addressed below. It is commonly believed that fistulectomy (excising the fistula) is more likely to result in incontinence than fistulotomy (laying the fistula open). Only two studies have randomized patients to these two procedures in order to resolve this issue. In the first, by Kronborg, 21 patients were randomized

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1147

to fistulectomy and 26 to fistulotomy [33]. Recurrence was similar in the two groups, occurring in 2 of 21 patients in the first and 3 of 24 in the second. Incontinence to flatus was rare in both groups, occurring in only 3 patients with fistulectomy and 1 with fistulotomy. Belmonte-Montes and coworkers also randomized patients to these two operations, but sadly presented no clinical outcome data in his publication [34]. The end point of that study was the ultrasonic measure of the length of sphincter division, demonstrating that patients having fistulectomy had in fact more sphincter divided by the technique employed by these authors than those having fistulotomy. Of course it would have been most interesting to see other end points presented in that paper, including time to healing, recurrence rate, and, most important, rate of incontinence. There are two other studies that specifically randomized patients with anal fistula. One assessed the role of marsupialization after fistulotomy; that is, suturing the skin edges to the fistula track to accelerate healing. Ho et al randomized 51 patients to marsupialization and 52 to traditional fistulotomy [35]. Healing time was 40% quicker in the marsupialized group and, interestingly, only 1 patient in that group developed minor incontinence postoperatively, whereas 6 members of the fistulotomy group developed incontinence. It was also stated that 4 out of these 7 patients had complete resolution of their incontinence with the passage of time. This suggests that what duration of wound healing may be as important as the length of sphincter division in the assessment of postoperative incontinence. The next RCT resurrected a technique first described 3000 years ago by Sushruta for the treatment of anal fistula [1], using a medicated Ayurvedic thread as a temporary draining seton in patients with fistula compared with those having fistulectomy. It was a large trial, comparing 265 patients treated with thread to 237 with fistulectomy [36]. Follow-up at one year was a problem. Only 60% of the patients could be followed. Incontinence was rare in both groups, though slightly less so in the thread group, and recurrence was equivalent in both groups, also quite rare. Rounding out the relevant RCTs (these are not directly related to key questions in the treatment of abscess and fistula treatment), Kausalya and Jacob compared epidural analgesia with general anesthetic in patients having anorectal surgery [37]. They found that the epidural was preferable because it allowed greater resting tone of the anal sphincter. Nessim and colleagues addressed the issue of whether postoperative bowel confinement was needed in a whole range of anal reconstructive procedures including fistula surgery and found it to be of no value [38]. One nonrandomized report having to do with anorectal abscess and fistula in very small infants is of some interest. Infants and smaller children with anal fissure are treated conservatively and rarely find their way into the operating room. This same conservatism was applied to small babies with abscess and fistula who were treated with drainage of abscesses or expectant treatment of established fistulas. All 18 patients examined had resolution of

1148

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

their fistulae spontaneously without surgical correction [39]. This suggests that fistulae in infants are a very different disease than fistulae in adults [40].

What we do not know The biggest problem facing a surgeon when addressing a patient with anal fistula is how to eradicate the problem without causing incontinence. This was once thought to be a minor problem. However, even some quite old publications demonstrate substantial risk of incontinence after fistulotomy, with incidence reaching as high as 55% for patients with horseshoe abscesses [41]. If the real and clinically significant prevalence of defecation dysfunction is as high as some reports suggest, it is a wonder that operative procedures for anal fistula have stood the test of time. Yet it seems that the vast majority of patients having fistula surgery feel satisfied with their procedure. In many reports of fistula surgery the reported incontinence rate is zero. One recent presentation described patients as essentially never voluntarily reporting incontinence in a surgeon’s office. When specifically asked if they were incontinent, 10% reported incontinence after fistula surgery. When the reporting of incontinence was confidential in the same patients through a mailed questionnaire, the rate shot up to 50% [42]. Clearly something is wrong with the way we assess postoperative continence in patients who are having anorectal procedures in general and treatment of abscess and fistula in particular. Is there a gold standard for assessment of postoperative incontinence? I think not. Is there a way to solve this problem? Yes, but not by attempting to report an absolute risk of incontinence. What is needed instead is reporting of the relative risk of incontinence, comparing several procedures in a properly conducted RCT. A standard continence assessment instrument and, most important, a uniform timing of continence assessment are needed in future studies. Related to abscess and fistula, a control group must constitute an accepted gold standard. For abscess, this seems to be unroofing and packing; for fistula, fistulotomy. There are new procedures proposed specifically to diminish the risk of incontinence in patients with fistula, but the usefulness of these procedures has yet to be rigorously established. Seton placement; that is, a stitch placed through the fistula track that is then either gradually tightened or left loosely in place and fistula divided in multiple stages, was first proposed by Hippocrates over 2500 years ago. It is still common practice in both forms [43]. RCTs have been presented at meetings, but never published, so the data could not be rigorously assessed. Newer procedures, such as closure of the internal opening of the fistula with either a mobilized flap of rectal mucosa [44] or anoderm [45], or the injection of fibrin glue [46] all appear to be attractive means of avoiding sphincter division in the treatment of fistula. Measurable success rates have been reported with these procedures, but their usefulness compared with fistulotomy in

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1149

eradicating the fistula, in preventing subsequent sepsis, and preserving continence is unknown.

Loose ends The relative efficacy and morbidity of fistulectomy versus fistulotomy have not been adequately resolved. We still don’t know whether patients with abscess and fistula should have definitive treatment at the time of abscess drainage. This second point may becoming moot in the United States as fewer and fewer patients are going to the operating room for their abscess drainage. Thus patients are more frequently getting definitive fistula therapy at a time remote from drainage, and do not go to surgery at all if the fistula never develops. Finally, it would be useful to revisit the rather interesting concept of primary abscess obliteration and closure. Early studies seem to favor this procedure. Its disappearance is puzzling. It is possible that it might gain broader acceptance if a less punitive suture technique were used.

References [1] Sankaran PS. Sushruta’s contribution to surgery. Varanasi: Indological Book House; 1976. [2] Ani An. Anorectal diseases in western Nigerian adults. A field study. Dis Colon Rectum 1983;26:381–5. [3] Navruzov SN, Dul’tsev IV, Salamov KN. Causes and prevention of rectal fistula recurrences. Vestn Khir Im I I Grek 1981;27:43–6. [4] Akinola DO, Hamed AD. Fistula in ano in Nigerians. Trop Gastroenterol 1989;10:153–7. [5] Read D, Abcarian H. A prospective survey of 474 patients with anorectal abscesses. Dis Colon Rectum 1979;22:566–8. [6] Ramanujam PS, Prasad ML, Abcarian H, et al. Perianal abscess and fistulas; a study of 1023 patients. Dis Colon Rectum 1984;27:593–7. [7] Sainio P. Fistula in ano in a defined population. Incidence and epidemiologic aspects. Ann Chir Gynaecol 1984;73:219–24. [8] http://www.cdc.gov. [9] http://hcup.ahrq.gov/HCUPnet.asp. [10] Vasilevsky CA, Gordon PH. The incidence of recurrent abscesses or fistula in ano following anorectal supperation. Dis Colon Rectum 1984;27:126–30. [11] Henrichsen S, Christensen J. Incidence of fistula in ano complicating anorectal sepsis: a prosepctive study. Br J Surg 1986;73:371–2. [12] Goligher JC. Surgery of the anus, rectum and colon. London: Balliere Tindall; 1975. [13] Nelson RL, Prasad L, Abcarian H. Anal carcinoma presenting as a perirectal abscess or fistula. Arch Surg 1985;120:632–5. [14] Nelson RL, Abcarian H. Do hemorrhoids cause cancer? Sem Colon & Rectal Surg 1995; 6:178–81. [15] Tarazi R, Nelson RL. Adenocarcinoma of the Anus. Sem in Colon & Rectal Surg 1995; 6:169–73. [16] Shacter E, Weitzman SA. Chronic inflammation and cancer. Oncology 2002;16:217–36. [17] Kuijpers HC, Schulpen T. Fistulography for fistula-in-ano. Is it useful? Dis Colon Rectum 1985;28(2):103–4.

1150

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

[18] Nelson RL, Schwartz A, Pavel D. Assessment of the usefulness of a diagnostic test: patient preference in diagnostic tests of colonic inflammation. BMC Med Res Methodol 2001;1:5. [19] Macfie J, Harvey J. The treatment of acute superficial abscesses: a prospective clinical trial. Br J Surg 1977;64:264–6. [20] Murie JA. The treatment of acute superficial abscesses; a prospective clinical trial. J Coll Surg Edinb 1981;26:282–5. [21] Lundhus E, Gjoede P, Gottrup F, et al. Bactericidal antimicrobial cover in primary suture of perianal or pilonidal abscess; a prospective, randomized, double blind clinical trial. Acta Chir Scand 1989;155:351–4. [22] Lundhus E, Gottrup F. Outcome at three to five years of primary closure of perianal and pilonidal abscess; a randomised, double blind clinical trial. Eur J Surg 1993;159:555–8. [23] Mortensen J, Kraglund K, Klaerke M, et al. Primary suture of anorectal abscess; a randomized study comparing treatment with Clindamyciin vs. Clindamycin and Getacoll. Dis Colon Rectum 1995;38:398–401. [24] Leaper DJ, Page RE, Rosenberg IL, et al. A controlled study comparing the conventional treatment of idiopathic anorectal abscess with that of incision, curettage and primary suture under systemic antibiotic cover. Dis Colon Rectum 1976;19:46–51. [25] Christensen AK, Kronborg O, Olsen H. Primaer sutur med eller uden daekke af klindamycin ved anorectal ae abscesser. Ugeskr Laeger 1983;145:576–8. [26] Kronborg O, Olsen H. Incision and drainage v. incision, curretage and suture under antibiotic cover in anorectal abscess; a randomized study with 3-year follow-up. Acta Chir Scand 1984;150:689–92. [27] Hebjoern M, Olsen O, Haakansson T, et al. A randomized trial of fistulotomy in perianal abscess. Scand J Gastroenterol 1987;22:174–6. [28] Schouten WR, van Vroonhoven TJMV. Treatment of anorectal abscess with or without primary fistulectomy; results of a prospective randomized trial. Dis Colon Rectum 1991; 34:60–3. [29] Nelson RL, Norton N, Cautley E, et al. The prevalence of fecal incontinence in Wisconsin households. JAMA 1995;274:559–62. [30] Tang CL, Chew SP, Seow-Choen F. Prospective randomized trial of drainage alone vs. drainage and fistulotomy for acute perianal abscess with proven internal opening. Dis Colon Rectum 1996;39:1415–7. [31] Ho YH, Tan M, Chui CH, et al. Randomized controlled trial of primary fistulotomy with drainage alone for perianal abscess. Dis Colon Rectum 1997;40:1435–8. [32] Li D, Yu B. Primary curative incision in the treatment of perianorectal abscess. Zhonghua Wai Ke Za zhi [Chinese Journal of Surgery] 1997;35:539–40. [33] Kronborg O. To lay open or excise a fistula in ano: a randomized trial. Br J Surg 1985; 72:970. [34] Belmonte-Montes C, Ruiz Galindo GH, Montes Villalobos JL, et al. Fistulotomia vs fistulectomia. Valoracion ultrasonografica de lesion al mecanismo de esfinter anal. Rev Gastroenterol Mex 1999;64:167–70. [35] Ho YH, Tan M, Leong AFPK, et al. Marsupialization of fistulotomy wounds imroves healing: a randomized controlled trial. Br J Surg 1998;85:105–7. [36] Shukla NK. Collaborating Centres ICMR. Multicentric randomized controlled trial of Kshaarasootra (Auyurvedic medicated thread) in the management of fistula in ano. Indian J Med Res 1991;94:177–85. [37] Kausalya R, Jacob R. Efficacy of low-dose epidural anaesthesia in surgery of the anal canal – a randomized controlled trial. Anaesth Intensive Care 1994;22:161–4. [38] Nessim A, Wexner SD, Agachan F, et al. Is bowel confinement necessary after anorectal reconstructive surgery? Dis Colon Rectum 1999;42:16–23. [39] Rosen NG, Gibbs DL, Soffer SZ, et al. The nonoperative treatment of fistula in ano. J Pediatr Surg 2000;35:938–9.

R. Nelson / Surg Clin N Am 82 (2002) 1139–1151

1151

[40] Fitzgeral RJ, Harding B, Ryan W. Fistula in ano in childhood: a congenital etiology. J Pediatr Surg 1985;20:80–1. [41] Bennett RC. A review of the results of conventional treatment of anal fistulae. J R Soc Med 1962;55:756–8. [42] Joy H, Williams JG. The outcome of surgery for high anal fistulas. Int J Colorectal Dis 2000;2:17. [43] Garcia-Aguilar J, Belmonte C, Wong DW, et al. Cutting seton versus two stage fistulotomy in the surgical management of high anal fistula. Br J Surg 1998;85:243–5. [44] Wedell J, Meier zu Essen P, Banzhaf G, et al. Sliding flap advancement for the treatment of high level fistulae. Br J Surg 1987;74:390–391. [45] Nelson RL, Cintron J, Abcarian H. Dermal island flap anoplasty for trans-sphincteric fistula in ano: assessment of treatment failures. Dis Colon Rectum 2000;43:681–4. [46] Cintron JR, Park JJ, Orsay CP, et al. Repair of Anorectal Fistulae with Fibrin Sealant-A Long-term Follow-up. Dis Colon Rectum 2000;43:944–50.