Fistula development after anal abscess drainage—a multicentre retrospective cohort study

Anal abscesses and fistulas represent acute and chronic manifestations of an infected anal gland [1, 2]. These conditions have been recognised and treated since antiquity, and symptoms include daily pain and discharge of blood and/or pus. Studies have shown decreased quality of life in patients with recurring anal abscesses and/or fistulas [3].

The treatment of both acute anal abscess and anal fistulas ranges from straightforward standardised incision and drainage to very complex procedures, depending on their anatomical location. Over the years, a plethora of methods and techniques for managing complex fistulas has been added to the surgical repertoire [4].

The development of fistulas in relation to acute anal abscess is well known, with rates ranging from 26 to 46% [5‐8]. However, little is known about the association between fistula development and the initial location of anal abscesses [9]. Endoanal ultrasound (EAUS) has joined MRI of the rectum and anal canal as the gold standard for evaluating anal sphincter pathology and is used for diagnosing and classifying both abscesses and fistulas [10, 11].

Clinical factors that could risk stratify patients with an abscess according to the risk of presenting with an anal fistula at follow-up have been assessed in earlier studies [6, 8, 12, 13]. One-quarter of all patients with Crohn’s disease suffer from Crohn’s-related perianal disease, including anal abscesses and fistulas. This is believed to have a different pathogenesis than cryptoglandular abscess and fistula diseases [9, 14]. Postoperative antibiotic treatment is believed to lower the risk of later fistulising disease, whereas patients with diabetes mellitus or gut-specific bacteria in pus cultures might have a higher risk [9, 15, 16]. On the other hand, neither age, sex or smoking habits have yet been associated with the development of perianal fistulas following an anal abscess [9, 17].

To date, Crohn’s has been the only known clinicopathological factor associated with fistula development following anal abscesses. To improve the treatment of patients with cryptoglandular anal fistulas, such as selection for follow-up, and thereby potentially early diagnosis and intervention, it is essential to identify risk factors for fistula development. Thus, this study aimed to identify the clinical risk factors for developing anal fistulas following surgically treated anal abscess disease.

This was a multicentre, retrospective cohort study of consecutive adult patients undergoing acute surgery for anal abscesses at two university hospitals in the Capital Region of Denmark in 2018 and 2019. Patients were identified using ICD-10-CM codes for anal abscesses, these being K61.0 to K61.5. Inclusion criteria were all patients above the age of 16 undergoing acute surgery for an anal abscess in the defined period of time. Patients with active fistula with/without seton were excluded. Data on predefined clinicopathological factors were extracted from patient records. These data included surgical notes describing the anatomical location of the abscess, suspicion of fistula, and surgical method, including the use of EAUS perioperatively. Data on specific patient histories were noted, most importantly, a history of anal fistula disease, recurrent abscesses, and inflammatory bowel disease (IBD). Furthermore, data on patient demographics, postoperative course, follow-up, other comorbidities, microbiology including pus cultures, and blood samples were extracted from the patient admission files.

This study was conducted in accordance with the STROBE guidelines for reporting cohort studies [18].

In Denmark, patients suspected for an anal abscess are referred to one of the surgical wards of the country. Clinical examination and, if needed, relevant imaging such as EAUS, CT or MRI are performed, and if anal abscess is suspected, the patient is booked for acute surgery. The abscess is drained according to its anatomical localisation and spread in the perianal spatias according to the skeletal muscle rule [19]. This states that all abscesses that penetrate the external anal sphincter (EAS) or levator ani muscle are to be drained percutaneously. If the abscess is localised medially to the EAS or levator ani, the drainage is performed endoluminally to the anal canal or rectum [20]. In cases of recurrent perianal abscess or fistula suspicion during surgery, the patient is directly referred to specialised follow-up 6 to 8 weeks post surgery. At discharge, patients are informed that non-healing of the wound after 4 weeks warrants further follow-up and, in such cases, they should be referred by their general practitioner. Clinical follow-up is performed by specialised proctologists with training in EAUS. The fistula diagnosis at follow-up is confirmed by EAUS in all cases and supplemental MRI in some cases. In case of fistula diagnosis, the patient is always offered further treatment depending on type and comorbidities. In our experience fistulas diagnosed at follow-up do not spontaneously heal, but need surgical intervention.

A total of 723 patients were identified using the ICD-10 codes. Ultimately, 248 patients were excluded because the date of surgery was outside the period of inclusion, they did not undergo surgery, or they had misdiagnosed abscesses (e.g. pilonidal or skin/nates abscesses). Finally, 475 patients were included in the final analysis (Fig. 1). This comprised 475 of 11,312 (4.2%) acute surgeries in the two hospitals during the inclusion period.

In total, 322 of the patients were males (67.8%), median age at surgery was 45 years (IQR 33–58 years) and median BMI was 26.1 kg/m² (IQR 23.0–30.2 kg/m²). A history of anal fistula disease was present in 37 (7.8%) patients and 121 (25.5%) had a history of recurrent anal abscess disease. Finally, 25 patients (5.3%) had Crohn’s (Table 1).

The present study identified several clinical risk factors that were statistically associated with the diagnosis of anal fistula following anal abscess incision.

At follow-up, 34.5% of the patients undergoing acute surgery for an anal abscess had a fistula. Earlier studies have reported similar rates [5, 6, 17].

Based on the anatomical localisation of anal abscesses, fistula development was significantly associated with low intersphincteric and ischioanal abscesses. Knowing the complexity and nature of high intersphincteric and supralevator abscesses, they are also expected to be associated with complex fistula disease. However, due to their rarity, only a few of these abscesses were included in this study. In a large retrospective cohort comprising 158,713 patients, Sahnan et al. determined the incidence of acute anal abscesses in England [17]. They found that ischioanal and intersphincteric abscesses were associated with fistula formation. However, the intersphincteric abscesses were not differentiated into low/high, and supralevator abscesses were not considered.

As expected, IBD was a significant risk factor for subsequent fistula formation. Almost 80% of the patients with Crohn’s disease or ulcerative colitis (UC) developed anal fistulas. Only four of the 29 patients with IBD had UC, while the remainder had Crohn’s disease. In the study by Sahnan et al., half of the patients with Crohn’s disease developed a fistula after first time admission with an acute anal abscess [17]. This supports the notion that anal abscesses in patients with Crohn’s are likely to originate from or progress into a perianal fistula.

CRP count of 100 mg/L or above on admission to the hospital was significantly associated with a persistent fistula at follow-up. CRP levels are known to increase in response to injury, infection and inflammation [21]. Most of these patients were, not surprisingly, diagnosed with complicated abscesses classified as high intersphincteric, ischioanal, horseshoe and supralevator—but few of them were treated with a seton at primary surgery. Interestingly, an increased white blood cell count was not associated with fistula development. Repeat surgery is fairly frequent when managing anal abscesses, although an association with fistula development has never been detected [22]. In this study, repeat surgery was highly associated with persistent fistulas at the follow-up. This was either a planned revision at the primary surgery or insufficient drainage was detected during the first postoperative days. Of all repeat surgeries, 37.8% had a CRP level > 100 mg/L upon hospital admission. This accounted for half of the patients with a CRP level of > 100 mg/L in the cohort and could support the notion that severe inflammation and reoperation are valid predictors of fistula formation.

Positive pus cultures with E. coli growth were also significantly associated with persistent fistulas at follow-up. In a retrospective analysis of patients with acute anal abscesses, Alabbad et al. reviewed the microbiological data of 211 patients with anal abscesses and found E. coli to be the most commonly isolated microorganism. In the present study, the most commonly isolated microorganism was also E. coli [23]. To our knowledge, no previous study has reported an association with later fistula development [12, 24]. However, the question remains whether it is the bacteria that cause fistulation to persist or whether it is a finding from an established fistula with acute infection.

Unsurprisingly, a history of recurrent anal abscesses was associated with persistent fistula at follow-up. Considering the pathophysiology of the abscess-fistula continuum of acute and chronic inflammation and the evolution of an abscess into a fistula, this makes sense [25]. In this study, one-fifth of the patients with acute abscesses had a fistula diagnosed during primary surgery. Half of these patients were treated only with abscess drainage (no seton or acute lay-open), and at follow-up, approximately one-fifth of them did not have a fistula and were spontaneously healed. When a fistula was only suspected during the primary surgery, only half of the patients were diagnosed with a fistula at follow-up. This could partly be the result of inexperienced surgeons examining the patients and performing the incision, but it also raises the question of how many fistulas heal spontaneously after acute drainage of an anal abscess.

The performance of EAUS during the primary surgery was significantly associated with anal fistula at follow-up. Considering these patients alone, the distribution of the anatomical localisation of the abscesses changed. The proportion of abscesses found in the perianal space was halved, whereas both low and high intersphincteric abscesses more than doubled. This likely represents a selection bias towards the use of EAUS in more complex patients in whom a fistula, complex abscess, or more difficult anatomy is suspected. In everyday clinical practice, this often leads to the procedure being performed by more experienced surgeons with better knowledge of anal anatomy, possibly explaining the change in abscess distribution. Of the patients who underwent EAUS during primary surgery, 44.4% were referred for follow-up and 75.9% of the referred patients had a fistula. Patients suspected of having complex abscess or fistula prior to surgery may have been selected for perioperative EAUS by more experienced surgeons. Thus, occult fistulas are detected at the time of abscess drainage, and many can persist until follow-up. This strengthens the role of ultrasound diagnostics not only in fistula disease in an elective setting but also in patients admitted with acute anal abscesses. In two thorough and recent statements from Italy and Germany, Amato et al. and Ommer et al. reported that 3D-EAUS and MRI, used in combination or alone, were of high and equal quality for diagnosing anal abscesses and fistulas [11, 26].

In this study, most of the abscesses drained endoluminally and, therefore, often guided by EAUS, were classified as high intersphincteric, supralevator, or with an intersphincteric/supralevator horseshoe formation. These anatomical localisations of abscesses are complex and, if not handled by an experienced surgeon with EAUS, lead to a risk of iatrogenic fistula formation (Fig. 4) [25‐28]. Patients not selected for EAUS in the acute setting are often managed by junior staff without sufficient experience in correctly classifying abscesses based solely on physical examinations. This could also affect the type of drainage performed, especially endoanal drainage, which may lead to iatrogenic damage and fistula formation.

The authors acknowledge the limitations of the present study. As with other retrospective studies, this study can only demonstrate associations, not causality. Missing data were inevitable and contributed as an essential limitation of this study. Another important limitation was the quality of the data regarding the anatomical localisation of the abscesses. Patient records were scrutinised for these data, but were often limited by less standardised or poorly specified operation notes. This can lead to incorrectly classified abscesses. As described by Sahnan et al., perianal abscess incisions are often performed by junior doctors, and thus insufficient knowledge of perianal spatias and anatomy is a common problem, and “perianal” is used as an umbrella term in many abscess cases instead of more detailed classifications like “ischioanal, low intersphincteric”. A sampling bias was also considered in this study. Patients were recruited from two hospitals in the same region of one country. This may have affected the generalisability of the results.