Introduction

Perianal abscess and anal fistulas are the acute and chronic manifestations of perianal infection [1]–[3]. It is estimated one-third of patients with perianal abscess may be complicate by chronic anal fistula formation [2], [3]. Perianal infections are often accompanied by severe pain, swelling, constipation, and may cause systemic infection and life-threatening sepsis [4], [5].

Infection is one of the most serious complications in patients with acute leukemia receiving chemotherapy [6], [7]. Perianal infection is not uncommon in patients with hematological malignancy [8]–[14]. Leukemic patients receiving chemotherapy and radiotherapy with resultant anal mucosa damage can provide an essential port of entry for pathogens during nadir stage [8]–[14]. The Disease spectrum of perianal infection might range from mild local cellulitis to life-threatening sepsis in patients with chemotherapy related neutropenia. Furthermore, neutropenic patients are unable to mount an adequate inflammatory response and are susceptible to infection by less virulent micro-organisms, consequently they may have unusual presentations of infection [15]. Not only can perianal infection cause mortality, but anal pain and discomfort also significantly impact the quality of life in patients with acute leukemia [14].

The principal management of perianal abscess and fistulas is surgery in the general population [1], [2]. Since leukemic patients are usually neutropenic with bleeding tendencies, surgical intervention could cause sepsis and poor wound healing [13]. Therefore, the optimal time for surgical treatment is still questionable in leukemic patients with perianal infection. To understand the prevalence, disease manifestations and outcomes of perianal infection in modern care, we retrospectively reviewed the medical records of 1102 acute leukemic patients in the National Taiwan University Hospital (NTUH) between 2001 and 2010 and analyze the clinical characteristic, microbiology and outcomes of perianal infection in acute leukemic adult patients.

Patients and Methods

Results

Epidemiology and clinical characteristics

During the study period, a total of 74 (6.7%) of 1102 adult patients with acute leukemia, included 59 (6.9%) of 853 patients with acute myeloid leukemia (AML) and 15 (6.0%) of 249 patients with acute lymphoblastic leukemia (ALL) were diagnosed with perianal infections. There were 57 men and 17 women with a median age of 43 years (range 18–77 years). The clinical characteristics were shown in Table 1. Sixty-six patients had perianal abscess, 30 patients had anal fistula infection, and 22 patients had both perianal abscess and fistula. Perianal infections were the initial presentation of acute leukemia in 14 of 74 patients (19%). Elderly patients (age ≧65 years) had less recurrent perianal infections than those with age <65 years (p = 0.05). AML patients had higher rates of recurrent perianal infections than ALL patients (p = 0.028).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Clinical characteristics of 74 adult leukemic patients with perianal abscess and/or fistula infection.

https://doi.org/10.1371/journal.pone.0060624.t001

Microbiology

The pathogens isolated from pus cultures and blood cultures were shown in Table 2. Overall, 116 and 25 microorganisms were isolated from pus and blood, respectively, from sixty patients. Forty-one (68%) of 60 patients had polymicrobial infections (2 to 5 pathogens). Gram-negative bacilli were the most predominant isolates (53% of pus culture isolates and 63% of blood culture isolates), followed by Gram-positive cocci, anaerobes and Candida (31%, 15% and 1% of pus culture isolates, and 25%, 8% and 1% of blood culture isolates, respectively). E. coli (25%) was the most common isolate, followed by Enterococcus species (22%), Klebsiella pneumoniae (13%) and Bacteroides species (11%). Three (12%) of 25 Enterococcus species isolated possessed vancomycin resistance, and 3 (10%) of 29 E. coli produced extended spectrum beta-lactamases (ESBL) in the pus culture.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Microbiology of 74 adult leukemic patients with perianal abscess and fistula infection.

https://doi.org/10.1371/journal.pone.0060624.t002

Treatment and outcomes

The prescribed antimicrobials for perianal abscess and anal fistula were heterogeneous, including 17 patients treated with carbapenem alone, 9 patients treated with carbapenem and vancomycin, 2 patients treated with carbapenem and piperacillin/tazobactam, 10 patients treated with cefepime alone, 9 patients treated with cefepime and metronidazole, 5 patients treated with cefepime and vancomycin; 9 patients treated with piperacillin/tazobactam alone; 3 patients treated with ticarcillin and clavulanate and aminoglycoside, one patient treated with ticarcillin/clavulanate and vancomycin, three patients treated with amoxicillin/clavulanate and aminoglycoside, and 5 patients treated with second generation or third generation cephalosporin and aminoglycoside. Twelve patients received a concomitant antifungal agent; 10 patients received fluconazole and 2 patients, amphotericin B. Whether the empirical antibiotics appropriately covered the isolated pathogens within 48 hour or not (early discordant antimicrobial therapy, 6.8% vs. 5.9%, p = 0.917) and coverage of anaerobic pathogens by either the initial or definitive antibiotics did not influence the 30 day outcome (8.1% vs. 0%, p = 0.315).

Twenty-five (34%) of 74 patients had undergone surgical intervention. Nineteen patients received incision and drainage, three patients received fistulectomy, two patients received fistulotomy, and one patient whose perianal abscess was complicated with Fournier's gangrene received debridement and fasciotomy. Four patients received two or three times incision and drainage. Five (20%) patients developed complications after surgical intervention, including four patients with sepsis after surgical intervention and one patient with poor wound healing. Nineteen (29%) of 65 neutropenic patients and 6 (67%) of 9 non-neutropenic patients received surgical intervention, neutropenic patients were more likely to not receive surgical intervention than non-neutropenic patients (p = 0.054). Surgical intervention for acute leukemic patients with anal fistulas showed a trend towards fewer recurrence (p = 0.067), but surgical intervention did not appear to decrease recurrences of perianal abscesses. The clinical characteristics of patients receiving surgery compared to those managed conservatively were no different in terms of age, gender, leukemia type, transplantation, and so on. Only three patients had magnetic resonance imaging to evaluate the anal lesion, including the patient with Fournier's gangrene. He had magnetic resonance imaging to evaluate the extent of lesion before surgery. Two other patients had ano-vaginal fistula. These two patients had relapsed or refractory leukemia, so they did not receive surgical intervention.

Univariate analysis of 30-day survival was performed with the Kaplan-Meier method and log-rank test (Table 3). Elderly (p = 0.015) and patients with shock status (p<0.001) had worse outcome. Multivariate analysis of 30-day outcome was performed with Cox regression analysis. Shock is the only independent factor predictive of 30-day crude mortality in hematologic patients with perianal infections (p = 0.016).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Univariate and multivariate analysis of 30-day outcome of 74 patients with perianal abscess and anal fistula infection.

https://doi.org/10.1371/journal.pone.0060624.t003

Discussion

Although the supportive management of patients has considerably improved in the past several decades, the prevalence of perianal infections in patients with acute leukemia of our study (6.7%) is similar to the result of previous studies (5∼9%) [8], [9], [15]. Even this disease entity is not common, perianal abscess and anal fistulas have long existed as problems for patients with acute leukemia, impacting the morbidity, mortality, and quality of life of these vulnerable patients.

The type of leukemia did not appreciably alter the prevalence of perianal abscess and/or anal fistula. Perianal infections usually recur in the period of neutropenia after chemotherapy. Patients with acute myeloid leukemia generally receive high dose cytarabine based chemotherapy with significantly worse neutropenia and mucositis than patients with acute lymphoblastic leukemia. This may explain why patients with acute myeloid leukemia had higher recurrence rate of perianal infection than patients with acute lymphoblastic leukemia (p = 0.028). Patients who were aged above 65 years had a trend towards lower recurrence rate than the younger patients. This lower recurrence rate could be an artifact related to the shorter overall survival of elderly patients. A larger scale study may be needed to confirm the findings of our relatively small single center study.

Few investigations have been held on the microbiology of perianal abscess and anal fistula infection in acute leukemia patients [8], [9]. Vanhueverzwyn et al. and Barnes and colleagues found Pseudomonas aeruginosa and E. coli were the predominant pathogens [8], [9]. In contrast, we found E. coli, Enterococcus and Klebsiella pneumoniae were the leading pathogens. Pseudomonas aeruginosa comprised only two percent of the cultured isolates. Although febrile neutropenia caused by Bacteroides species were few (about 3%) [7], [16], Bacteroides species represented 10% overall of the isolated pathogens in this study. Multivariate analysis revealed shock as the independent factor associated with poor outcomes for these leukemic patients with perianal infection. The antimicrobial coverage of anaerobes could be better control of perianal infection in hematological patients. Liu et al and Eykyn et al. reported that E. coli and Bacteroides species were the leading pathogens in the general population [17], [18]. Enterococcus species occupied 22% of all pathogens isolated in the study. Leukemia patients are frequently given broad spectrum antibiotics for febrile neutropenia. Gram negative bacteria decreased under broad spectrum antibiotics whilst gram positive enterococcus increased. Tuberculosis and actinomycosis have also been reported to cause perianal infections among neutropenic patients [19], [20], but there were no cases found in this study. Our results reasonable suggest that Enterobacteriaceae are the most common causative pathogens of perianal infection, not streptococci or staphylocci, which are the most common causative pathogens of other site of skin and soft tissue infections.

The principal management of perianal abscess and anal fistula is surgery [1], [2]. Abscess should be drained in a timely manner. Delayed or inadequate treatment may cause extensive life-threatening infection and sepsis [1]. In this study, neither early discordant antimicrobial therapy nor inadequate antimicrobial coverage of anaerobic pathogens influence the 30-day mortality. The findings might reflect low mortality in this disease entity and effect of spontaneous pustular drainage or surgical debridement for better infection control. Meta-analysis by Malik and colleagues showed fistula surgery with abscess drainage significantly reduces recurrence or persistence of abscess fistula in general population [21]. The recurrence rate is 3 to 5% in the general population [17], [22]. The recurrence of perianal abscess and anal fistula infection is 31% in this study. This high recurrence rate underscores the importance of perianal disease for patients with acute leukemia. However, patients with acute leukemia commonly develop neutropenia and thrombocytopenia after chemotherapy. Moreover wound healing is impaired and sepsis may worsen after surgical intervention. The optimal time for surgical treatment remains a challenge for leukemic patients with perianal abscess and fistulas. Clinically, a surgeon may act conservatively if the patients' leukemia is not in remission. Some novel treatment such as plug in anal fistula may be an option for such leukemic patients [23], [24]. Recently, MRI exquisitely depicts the perianal anatomy and shows the fistulous tracks and their associated ramifications and abscesses [25], [26]. MRI appears to be a good tool to evaluate leukemic patients with perianal lesions, especially in the patients with complex anal fistula and complicated abscess.

There are some limitations to this study. First, the disease spectrum of perianal infection might range from mild local heat and tenderness to life threatening sepsis, especially in patients with hematologic malignancies received chemotherapy. In order to clearly define this issue, we limit the disease category specific to perianal abscess or anal fistula, which are more severe clinical condition of perianal infection. Therefore, the actual prevalence of perianal infection could be underestimated. Second, isolation of micro-organism from purulent discharge from an unsterile site is not perfect for identification of causative pathogens. In this retrospective study, we found that the pathogens derived from blood and discharge is similar (table 2). Therefore, we suggest the result of microbiology could represent partly the real pathogens of ano-rectal infection. Third, the detailed anatomic classifications of anal fistulas could not be obtained and there is no data to detail their miserable quality of life because of the retrospective study design.

Acknowledgments

We thank the staff of the Eighth Core Lab., Department of Medical Research, National Taiwan University Hospital, for their technical support during the study.