perineal infections caused by Pseudomonas aeruginosa in children with malignant diseases

Management Pseudomonas

of Anorectal/Perineal in Children

aeruginosa

Infections Caused by With Malignant Diseases

Carlos Angel, Christian C. Patrick, Thorn Lobe, Bhaskar Rao, and Ching-Hon Pui Memphis, l The role of operation for anorectal infections associated with perineal gangrene and cellulitis in children with myelosuppression from cancer chemotherapy is unclear. We evalu-

ated anorectal/perineal infections caused by Pseudomonas aeruginosa in 16 children with malignant diseases seen over 27 years. In 12 of 16 patients, leukemia was the underlying malignancy (ALL 10, AML 2). and in 13 of 16, severe neutropenia (absolute neutrophil count <500/mm’) was present at diagnosis. Cultures of the lesions showed multiple organisms in 14 of 16 patients with Escherichia co& Klebsiella species, and Enterococcus being the most frequent coexisting organisms. All positive blood cultures grew Paeruginosa exclusively. of three patients with necrotizing infections, two had complete resolution with medical treatment alone; the other patient who developed this problem while on terminal care died. In none of the 16 patients was a major operation (debridement or diversion) performed. Five patients died, three of whom were considered terminally ill when the anorectal infections occurred. Four of the five deaths occurred before 1974. Since then, only 1 of 7 patients died. Excluding the three terminally ill patients, the success rate of medical therapy alone is 65% (11/13). The antibiotic regimen should include an aminoglycoside in synergistic combination with anti-Pseudomonas penicillin. These results suggest that operative management may have no role in the management of anorectal infections caused by P aeruginosa in children with cancer. Copyright o 1997 by W.B. Saunders Company INDEX WORDS: Anorectal infections, Pseudomonas nosa; cancer chemotherapy.

aerugi-

H

anorectal infections in paISTORICALLY, tients with malignancies have been associated with a high mortality.‘-4 These infections may lead to necrotizing fasciitis and, when involving the scrotum and penis, to classic Fournier’s gangrene.‘Pseudomonas aeruginosa is the most frequent infective agent. In neutropenic patients the classical clinical findings of swelling and fluctuance appear late in the course of the infection or not at all, and there may be scant or no pus formation. Improvements in the supportive care of immunocompromised patients have greatly reduced the morbidity and mortality of this disease in recent years. Surgeons have frequently been reluctant to operate on these patients in the presence of low neutrophil and platelet counts, fearing complications such as hemorrhage, poor wound healing, and the spread of infection.‘j However, a recent study has shown that operations can be performed without a prohibitive complication rate.6 Nonetheless, whether

Journal ofPediafric Surgery, Vol26, No 4 (April), 1991: pp 487-493

Tennessee

operative management can improve the outcome of this disease remains a controversial issue.6,7 This report reviews our experience with anorectal infections caused by P aeruginosa in children with malignancies to establish the natural history of this entity and the results of treatment in this patient population. MATERIALS AND METHODS A retrospective, 27-year, computerized search of patients admitted for treatment of malignant diseases at St Jude Children’s Research Hospital (SJCRH) was carried out involving the discharge diagnoses of Pseudomonas cellulitis of the buttocks or perineum, perianal cellulitis or abscess, perirectal abscess, and/or perirectal ulcer. The criteria for patient inclusion were the presence of swelling, induration, erythema, pain, or ulcer formation in the perirectal area associated with a positive culture for P aeruginow. The patient’s age, sex, race, underlying malignancy, clinical and hematologic status, mode of presentation, time of presentation, and the characteristics of the lesion were recorded. Also tabulated were the administration of antibiotics, steroids, or other chemotherapeutic agents within 30 days before presentation of the lesion, results of laboratory studies at presentation (with particular emphasis on the absolute neutrophil count), culture results, and the organism’s sensitivities from lesions and blood, treatment modalities, outcomes, complications, and deaths. The following definitions were used: 1. Terminal patient: patient with progressive malignant disease, whose death was considered inevitable and imminent. 2. Cure: total resolution of the perineal infection with improvement of pain, defervescence of fever, negative blood cultures in patients with bacteremia, and complete epithelialization of the ulcer or presence of healthy granulation tissue with no further local growth of Pseudomonas. 3. Recurrence: reappearance of a rectal infectious lesion after cure.

From the Section of Pediatric Surgery, and the Departments of Infectious Diseases, Hematology-Oncology, and Pediatrics, St Jude Children S Research Hospital, LeBonheur Children S Medical Center, and The University of Tennessee, Memphis, TN. Supported in part by Childhood Solid Tumor Program Project Grant No. CA-23099 and Cancer Center Support (CORE) Grant No. CA-21 765from the National Cancer Institute, Bethesda, MD, and the American Lebanese Syrian Associated Charities (ALSAC). Presented at the 21st Annual Meeting of the American Pediatric Surgical Association, Vancouver, British Columbia, May 19-22, 1990. Address reprint requests to Thorn E. Lobe, MD, St Jude Children S Research Hospital, Depatiment of Surgery, 332 N Lauderdale, Memphis, TN 38101. Copyright o 1991 by WE. Saunders Company 0022-3468/9112604-0023$03.OOlO

487

488

ANGEL ET AL

4. Failure: persistence or progression of infection despite medical treatment.

RESULTS

Between 1962 and December 1, 1989, 5,747 children with malignancies were admitted to SJCRH. Of these, 131 (2.28%) patients developed anorectal/ perineal infectious lesions. The most commonly isolated species were Escherichia coli and Enterococcus. Paeruginosa was isolated from the lesions in 16 of the patients, for an incidence of 0.28% of patients admitted and 12.2% of the patients who develop an anorectal/perineal lesion. There were 5 males and 11 females. Fourteen patients were white. Their ages ranged from 9 months to 13 years 4 months (mean, 4.8 years). Most (12/16) of the patients had leukemia (ALL 10, AML 2). Also included were 2 patients with neuroblastoma and 1 each with osteosarcoma and rhabdomyosarcoma. Thirteen patients had active disease at diagnosis; 2 patients with ALL and 1 patient with osteosarcoma were free of malignant disease when their anorectal lesions were noted. The timing of presentation varied: in 11 patients, lesions occurred during or immediately after chemotherapy (5 occurred during induction chemotherapy), in 3 patients lesions occurred as part of a terminal event, and in only 2 patients lesions occurred synchronously with the first diagnosis of their malignancy. In 3 patients the anorectal lesions appeared in association with Pseudomonas sepsis with positive blood cultures for P aeruginosa being present several days before the development of the perianal lesions. The most frequent symptoms were localized pain, irritability, and fever (> 38.5”(Z), which were present in 14 of the patients. In only 4 patients did spontaneous drainage of the lesion(s) occur. Ten patients had lesions involving only the perirectal areas, and six patients had lesions that involved the perianal and genital areas. Most of the lesions were erythematous and indurated or ulcerated and only 3 patients had necrotizing infections that involved the perirectal and genital areas; 2 of these patients healed completely with medical treatment alone and 1, who developed this problem while on terminal care, died. Thirteen patients received chemotherapy in the 30 days before diagnosis. Seven patients had received steroids and 14 patients had received antibiotic treatment of febrile illnesses, mainly fever and sepsis of unknown origin. These antibiotics included antiPseudomonas coverage in 10 of the cases. No patient had radiotherapy in the month before developing anorectal lesions. Four of six patients who had perirectal and genital lesions had received steroids in

the month before their diagnosis compared with 3 of 10 who developed perirectal lesions only. The absolute neutrophil count (ANC) ranged from 0 to 1,680/mm3 (mean, 210/mm3); 13 patients had an ANC <500/mm3. The hematocrit ranged from 23% to 35% (mean, 28.3%). In 12 of 16 patients the platelet count was <50,000/mm3. In the 5 patients who died, the number of days of neutropenia before diagnosis ranged from 5 to 27 (mean, 17 days); in those who survived it ranged from 3 to 25 (mean, 12 days). However, the total days of neutropenia (including days of neutropenia before diagnosis and after) was shorter in the patients who died (2 to 29; mean, 13.2 days) than in patients who survived (0 to 58; mean, 17.7 days). All lesions were cultured and grew P aeruginosa. Fourteen cultures grew multiple organisms, with E coli, Klebsiella species, and Enterococcus being the most frequent coexisting organisms. Blood cultures were performed in 15 patients. Ten of those cultures had no growth, 4 cultures grew P aeruginosa alone, and 1 culture grew Staphylococcus epidermidis, a probable contaminant. The organisms obtained from blood cultures were sensitive to gentamicin (4/4), clindamycin (3/4), polymyxin B (3/4), and carbenicillin (2/4). Fifteen of 16 of the organisms isolated from the anorectal lesions were sensitive to an aminoglycoside, 11 of 16 were sensitive to polymyxin B, and 10 of 16 were sensitive to carbenicillin. Since 1980, the specimens have been tested showing sensitivity to amikacin (5/5), cefotaxime (4/4), ceftazidime (2/2), and ticarcillin (2/2). Six of seven specimens were sensitive to clindamycin. Local treatment consisted of sitz baths (halfstrength betadine or peroxide) or warm saline compresses several times a day. This was accompanied by a liberal application of a topical antimicrobial cream such as Sylvadine or Sulfamylon. On rare occasions sloughed, necrotic tissue was removed at the bedside. The patients were generally placed on a commercially available air mattress. No patient had operative intervention. Fifteen patients received an aminoglycoside either before or after diagnosis. Three of the five patients who died and 4 of 11 of those who survived had been treated with an aminoglycoside before diagnosis (Table 1). This did not prevent the appearance of perirectal lesions. After diagnosis, the most frequent antibiotics used were gentamicin (14/16), carbenicillin (8/16), and polymyxin B (7/16). The combination of gentamicin and carbenicillin was used in 9 patients and the combination of amikacin and ticarcillin was used in 3 patients. Eleven patients were successfully treated; 5 patients died. One patient’s lesion had resolved with treatment but recurred 23

ANORECTAL/PERINEAL

Table 1. Pseudomonas Case NO.

489

INFECTIONS IN MALIGNANCY

Age of Diagnosis Sex/Race (Year)

Malignancy (Statusat Dx)

Anorectal Infections in Children With Cancer-Suwivors

Antibiotics30 Days Priorto Dx

Chemotherapy30 DaYsPriorto Dx

Steroids30 Days Priorto Dx

Atea

Result (Complicationsl

1

2 yt, WM (1971)

AML (A)

No

No

No

Petitectal

Cute (anal fissure)

2

9mo,WF(l972)

RMS (A)

Ox, K, Cef, Gent, CB

VCR, Dactino, Cycle

No

Petitectai

Cute (none)

3

7% yt. WF (1973)

ALL (CR)

Pen, CL, G

No

No

Petitectal

& genital Cute (none)

&genital 4

3 yt 4 mo, WF (1973)

ALL (CR)

Amp

MTX, Cycle, Ata-C, 6MP

5

2 yt 1 mo, WM (1974)

ALL (A)

Cef, Gent

VCR, L-aspat

6

1 yt 5 mo, WF (1977)

ALL (A)

Pen, policillin

No

7

7 Yt 1 mo, WF (1979)

ALL (A)

Gent, Ox

Dauno, VCR

Petitectal

Cute (none)

Petitectal

Cute (none)

No

Petitectal

Cute (none)

Ptednisone

Petitectal

Cute (none)

No

&genital 8

13yt4mo,HF(1981)

ALL (A)

Ox, Gent, CB

VM-26, Ata-C, VCR,

Ptednisone

Petitectal

Cute (none)

L-Aspat 9 IO

IOyt, WM (1985)

NB (Al

ox, Cfx

VP-16

No

Petitectal

Cute (none)

3 yt IO mo, WM (1988)

ALL (A)

Cfx, Ox, Vane, CL

VCR, Dauno, MTX

Solumedtol

Petitectal

Cute (none)

& genital 11

8 yt 9 mo, HM (1988)

OS (CR)

ox, Cfx

CP, Adtia

No

Petitectal

Cute, recurred later (petitectal fissure)

Abbreviations: W, white; M, male; F,female; H, Hispanic; Dx, Diagnosis; NA, not available; NG, no growth; Gent, gentamicin; Strep, streptomycin; Ox, oxacillin; Cfz, ceftazidime; CB, catbenicillin; PB, polymyxin B; K, kanamycin; Cef, cephalothin; Pen, penicillin; CL, clindamycin; G, ganttisin; TMX, Ttimethoptim-sulfamethoxazole;

Cfx, cefotaxime; Vane, vancomycin; Amik, amikacin; Ticat, ticatcillin; AML, acute myelocytic leukemia; ALL,

acute lymphocytic leukemia; RMS. thabdomyosatcoma; dactinomycin; Cycle, cyclophosphamide;

OS, osteosatcoma;

MTX, methottexate;

NB, neutoblastoma; CR, complete remission; VCR, vinctistine; Dactino,

mAMSA, amsactine; L-Aspat, L-aspataginase; Adtia, adtiamycin: CP, &-platinum;

Ata-C, cytosine atabinoside-C; 6MP, 6-metcaptoputine.

days later after a subsequent course of chemotherapy. The culture of the recurrent lesion failed to show any bacterial growth and although the lesion resolved with a new course of anti-Pseudomonas therapy, a residual perirectal fistula occurred. Overall, few complications resulted from the anorectal infections in the patients who survived. Two patients developed perianal fistulas and 1 patient had scrotal and perianal gangrene that healed without sequelae. In the surviving patients, reduction of fever with antibiotic therapy after diagnosis ranged from 1 to 15 days (mean, 4.09 days) and the time until the ANC was above 500/mm3 ranged from 0 to 51 days (mean, 13 days). Overall, 5 of 16 patients died (Tables 2 and 3). All patients were white girls (aged 1 year 3 months to 4 years 8 months). Four of the five deaths occurred before 1974. Since then only 1 of 7 patients treated for Pseudomonas anorectal infections died. Three patients with ALL were in relapse, 2 of whom were considered terminal. One patient had disseminated neuroblastoma (also terminal) and 1 patient with AML died of Pseudomonas infection during induction chemotherapy. Only 2 patients had appropriate antiPseudomonas antibiotic coverage in the 30 days before the diagnosis of their anorectal infection and all patients had received chemotherapy during this interval. All patients were neutropenic at diagnosis. Blood cultures were done in 4 of the patients and 3 were positive for P aeruginosa. One of the terminal patients received no antibiotic treatment for her anorectal

infection;

the rest of the patients received antithat, in the case of another terminal patient, was discontinued at the parent’s request. The blood cultures at autopsy were positive for P aeruginosa in 4 of 5 patients. The interval from diagnosis to death ranged from 2 to 7 days (mean, 3.6 days). None of the patients in remission died; only 1 of 7 patients presenting with this infection at diagnosis or during induction chemotherapy died. The mortality of patients in relapse was high (4/5) and, discounting the expected deaths, overall mortality was 15.3%.

Pseudomonas treatment

DISCUSSION

P aeruginosa infections occur frequently in children and adults with cancer and can be responsible for their deaths.7-9 This series showed that in 12.2% of the children with cancer who develop an anorectal lesion, Pseudomonas was the causal agent. Most cases of anorectal lesions in patients with cancer in the literature have occurred in patients with leukemia, but patients with other malignancies may also develop anorectal lesions if they become neutropenic.‘.‘” Traditional risk factors for Pseudomonas infections in cancer patients have been neutropenia, concurrent chemotherapy, and recent treatment with antibiotics.l.8’” In this study, 13 of 16 patients were neutropenic at presentation, 12 of whom had received chemotherapy in the 30 days previous to diagnosis and 14 of 16 had received antibiotics in the 30 days prior to diagnosis. In contrast to adult cancer patients

490

ANGEL ET AL

Table 2. Pseudomonas Anorectal Infections in Children with Cancer--Survivors Case No.

Culture of Abscess

1

Pseudomonas

2

Pseudomonas

Blood

ANC

Total Days

Days of ANC

Culture

at Dx

of ANC < 500

< 500 Prior to Dx

NG

NA

NA

NA

Gent, Ox, Carb, PB (8 d)

42

16

11

Gent, CB, CL, PB (8 d)

NG

800

7

7

Gent, PB (7 d)

NG

1,680

0

0

PB, Gent, Cef, (5 d)

NG

150

10

5

CB, PB, Gent, Cef (6 d)

NG

46

15

NA

Treatment (Duration)

E coli Pseudomonas

E coli Klebsiella Enterococcus Proteus 3

Pseudomonas E coli Klebsiella Enterococcus

4

Pseudomonas Enterococcus

5

Pseudomonas E coli a Streptococcus

6

Pseudomonas

Ox, Gent, CB, Pi3 (18 d)

Enterococcus Proteus 7

Pseudomonas

NG

0

16

6

8

Pseudomonas

NG

30

28

25

Ox, Gent, CB, TMX, Chl, Vane, Col (14 d)

NG

0

21

13

Vane, Amik, Ticar, Col, TMX (8 d)

60

58

8

Cfx, Ox, Col, Vane, Amik, Ticar, TMX, CL (57 d)

90

6

3

Amik, Ticar, Ox (8 d)

Ox, CB, Gent, Chl, TMX (14 d)

E coli Klebsiella Pseudomonas E coli a Streptococcus Pseudomonas

Pseudomonas

E coli Klebsiella S epidermitis Pseudomonas

NG

E coli S epidermitis Abbreviations:

ANC, absolute neutrophil count; Dx, diagnosis;

NG, no growth; NA, not available; Gent, gentamicin;

carbenicillin: PB, polymyxin B; CB, carbenicillin; CL, clindamycin; Cef, cephalothin; Chl, chloramphenicol:

Ox, oxaciliin; Carb,

TMX, trimethoprim-sulfamethoxazole;

Vane, vancomycin; Col, colistin; Amik, amikacin; Cfx, cefotaxime; CL, clindamycin.

in whom the primary source of Pseudomonas infection is the respiratory tract, the most common portal of entry for Pseudomonas infections in children with malignancies is the gastrointestinal tract.‘,” In most cases no predisposing factor for infection is present. The infection may remain localized or spread into the ischiorectal fossa, the genital area, the supralevator space, peritoneum, or retroperitoneum. Seeding into the bloodstream may occur with localized or invasive lesions.’ If the scrotum and penis are involved, Fournier’s gangrene may result. This has been reported to occur at any age, including neonates.5,‘3 Many authors recommend early aggressive operative management of patients with Fournier’s syndrome.‘4-16 The only patient in this series with necrotizing infection involving the perineum, scrotum, and penis healed completely with aggressive medical treatment, which included granulocyte transfusions. In this series, 4 of 6 of the patients who developed

anorectal and genital lesions had received steroids within 30 days before their diagnosis compared with only 3 of 10 of those in whom the lesions remained localized to the perianal area. In the study by Barnes et aJ6 the perianal lesions appeared while the patient was hospitalized and were always preceded by severe neutropenia. This was true for most of the present patients, but two of them had anorectal lesions as part of the initial presentation of their malignancy and only 13 of 1.5 were neutropenic when their lesions appeared. The most frequent symptoms were fever and localized pain; spontaneous drainage occurred in only 4 of 16 patients, perhaps due to scant pus formation by neutropenic patients. Anemia, neutropenia, and thrombocytopenia were also frequent findings. Anorectal lesions can be the source of bacteremia or the result of Pseudomonas sepsis; 3 of the patients had positive blood cultures several days before the appearance of perianal lesions. Although mucositis due to

ANORECTAL/PERINEAL

491

INFECTIONS IN MALIGNANCY

Pseudomonas Anorectal Infections in Children With Cancer-Deaths

Table 3.

Patient 1

Patient 2

Patient 3

Patient 4

Patient 5

Age (vr)

2yrl

lyr7mo

4yr4mo

Race/sex

WF

WF

WF

WF

Year

1963

1967

1972

1974

1986

Malignancy

NB

AML

ALL

ALL

ALL

Status at Dx

Terminal

Active (during

Terminal

Terminal

mo

1

yr3mo

4yr8mo WF

induction) Antibiotics last 30 days

No

Ox, Strep

Amp, Ox

Gent, CL, Chl, Cef

Vane, Amik, Ticar,

No

No

Yes

Yes VM26, mAMSA

Chl, TMX Adequate for isolate Chemotherapy last 30 days

CYCIO

VCR, 6MP

VCR, 6MP

VCR, Ara-C

Steroids last 30 days

Prednisone

Prednisone

No

Prednisone

No

Culture of abscess

Pseudomonas

Pseudomonas

Pseudomonas

Pseudomonas

Pseudomonas

E co/i

E coli

E coli

E co/i S epidermitis

Enterococcus Proteus

Not done

Blood culture

NG

Pseudomonas

Pseudomonas

Pseudomonas

ANC at Dx

96

42

36

26

100

Total no. of days of ANC < 500

2

16

21

29

7

Treatment after diagnosis

Col, Pen

Gent, CB, Col, PB

Gent, CB, PB, Cef, Chl

Cef, Gent, CB (stopped

None

at parents request) Interval (d) from diagnosis

5

2

I(

2

7

L

to death Autopsy

Blood culture, Pseudomonas

Rectum and CSF, Pseudomonas; blood, Candida;

Blood culture, Pseudomonas

Blood culture, no growth

lung, Pffeumocystis Abbreviations:

Dx, diagnosis; ANC, absolute neutrophil count; W, white; F, female; NB, neuroblastoma;

Cycle, cyclophosphamide;

NG, no

growth; Col, colistin; Pen, penicillin; AML, acute myelocytic leukemia; Ox, oxacillin; Strep, streptomycin; VCR, vincristine; 6MP, 6-mercaptopurine; Gent, gentamicin; CB, carbenicillin; PB, polymyxin B; ALL, acute lymphocytic leukemia: Amp, ampicillin; Cef, cephalothin; Chl, chloramphenicol; CSF, cerebrospinal fluid; CL, clindamycin; Ara-C, cytosine arabinoside; Vane, vancomycin; Amik. amikacin; TMX, trimethoprom-sulfamethoxazole; VM26, etoposide; mAMSA, amsacrine.

chemotherapy can be an important etiologic factor in the appearance of anorectal lesions, only one fourth of the patients received agents such as cytosine arabinoside and methotrexate, which are particularly prone to cause mucositis. The administration of anti-Pseudomonas treatment in 10 of 14 patients 1 month before diagnosis did not prevent the development of the anorectal lesions. This was probably the result of an active malignancy, persistent neutropenia, and concurrent chemotherapy alone or in combination. Previous studies have shown the synergistic polybacterial nature of these infections.h.7.8~‘7 In this study, 14 of 16 of the cultures from the lesions grew multiple organisms. No cultures grew anaerobes, but specific anaerobic cultures were not requested and, therefore, the importance of the aerobic-anaerobic synergism could not be assessed.7 Bacteremia was present in only 4 of 15 of the patients who had blood cultures done and, in contrast to previous studies, multiple organisms were not isolated from the blood.@ There is controversy with respect to some areas of treatment. Although most agree that local care is needed to keep the lesions clean and to improve the patient’s comfort, there is conflicting evidence in the literature regarding the role of operative intervention

in the treatment of this condition. Barnes et aI6 advocate early aggressive operative drainage and debridement in the presence of persistent fever, neutropenia, failure of the lesions to drain spontaneously, progression of disease, genital involvement, and peritoneal signs. In contrast to that series in which more than 50% patients had necrotizing infections; only three of the present patients developed this type of infection. Two of them healed completely with aggressive medical treatment alone and the third was on terminal care when he became infected. Glenn et al7 found that when combining an aminoglycoside with an antibiotic specific for anaerobes, the cure rate was 88%, which was comparable to the results obtained when surgery was done. They did not feel that operation improved the outcome of this disease and recommended the combination of a third-generation cephalosporin with activity against Pseudomonas (ceftazidime) with clindamycin, or the combination of an extended-spectrum penicillin (piperacillin) with an aminoglycoside and clindamycin7 In cases of progressive disease, vancomycin could be added to maximize gram-positive coverage especially for enterococci. In 15 of 16 cases, the organisms recovered from patients were sensitive to an aminoglycoside, and the sensitivities of isolates in recent years

492

ANGEL ET AL

show the efficacy of third-generation cephalosporins such as ceftazidime and anti-Pseudomonas penicillins such as ticarcillin. Three fourths of the patients received treatment (after appearance of anorectal lesions) with a combination of either gentamicin and carbenicillin, or amikacin and ticarcillin. In only 1 patient was clindamycin added to the combination. Excluding the expected deaths of 3 terminal patients, the success rate of medical treatment alone in this series was 85%, which is comparable to the results of Glenn et al’ and confirms the efficacy of aggressive antibiotic therapy in these patients. However, we assume that the anaerobic coverage of anti-Pseudomonas penicillins is probably adequate because most of the patients healed without requiring specific medication against anaerobes. However, the addition of clindamycin is a reasonable option in patients with progressive disease. Granulocyte transfusions were used in the two patients who developed necrotizing infections. The use of granulocyte transfusions is at present controversial and should be limited

to severely neutropenic patients with a documented gram-negative infection who do not respond to antibiotic therapy and in whom prompt bone marrow recovery is unlikely.‘8~19 None of the present patients had operations as advocated by Barnes et ale6Most of the lesions healed completely, with only 2 resulting in perirectal fistulas. The clinical improvement of the lesions seemed to be associated with increasing neutrophil counts and not with a change in the status of the malignancy. The mortality of the present patients was associated mainly with the status of the malignancy and persistent neutropenia. All of the patients who died remained neutropenic until their deaths. Whereas 4 of 5 of the patients with relapsing malignant disease and anorectal lesions died, only 1 of 7 died during induction and none died when their disease was in complete remission at diagnosis. That 4 of 5 of the deaths occurred before 1974 perhaps reflects the improvements in supportive care for pediatric cancer patients that have taken place in recent years.

REFERENCES 1. Schimpff SC, Wiernik PH, Block JB: Rectal abscesses in cancer patients. Lancet 2:844-847,1972 2. Sehdev MK, Dowling MD Jr, Seal SH, et al: Perianal and anorectal complications in leukemia. Cancer 31:149-152,1973 3. Howard RJ: Nonspecific host defenses in surgical cancer patients. Curr Probl Cancer 7:1-39, 1983 4. Blank WA: Anorectal complications in leukemia. Am J Surg 90:738-741, 1955 5. Berg A, Armitage JO, Burns P: Fouriner’s gangrene complicating aggressive therapy for hematologic malignancy. Cancer 57:2291-2294,1986 6. Barnes S, Sattler F, Ballard JO: Perirectal infections in acute leukemia: Improved survival after incision and debridement. Ann Intern Med 100:515-518,1984 7. Glenn J, Cotton D, Wesley R, et al: Anorectal infections in patients with malignant disease. Rev Infect Dis 10:42-52,1988 8. Jackson MA, Wong KY, Lampkin B: Pseudomonas aeruginosa septicemia in pediatric cancer patients. Pediatr Infect Dis J 1:239-241, 1982 9. Earle MF, Fossieck BE Jr, Cohen MH, et al: Perirectal infections in patients with small cell lung cancer. JAMA 246:24642466.1981 10. Hagihari P, Howard R: Anal and perianal infections, in

Simmons R, Howard R (eds): Surgical Infectious Diseases. East Norwalk, CT, Appleton-Century-Crofts, 1982, pp 987-995 11. Bodey GP: Epidemiological studies of Pseudomonas sp in patients with leukemia. Am J Med Sci 260:62-65, 1970 12. Whitecar JP Jr, Luna M, Bodey GP: Pseudomonas bacteremia in patients with malignant disease. Am J Med Sci 260:216-223, 1970 13. Sussman SJ, Schiller RP, Shashikumar VL: Fouriner’s syndrome: Report of three cases and review of the literature. Am J Dis Child 132:1189-1191,1978 14. Spirnak JP, Resnick MI, Hampel N, et al: Fouriner’s gangrene: Report of 20 patients. J Urol 131:289-291,1984 1.5. Diettric NA, Mason JH: Fouriner’s gangrene: A general surgery problem. World J Surg 7:288-294,1983 16. Pollack EW, Frieden F, Ozar M: Periscrotal gangrene. Two staged therapeutic approach. South Med J 74:1040-1042,198l 17. Brook I, Marten WJ: Aerobic and anaerobic bacteriology of perirectal abscess in children. Pediatrics 66:282-284,198O 18. DiNubile M: Therapeutic role of granulocyte transfusions. Rev Infect Dis 7:232-243,1985 19. Strauss RJ: The role of granulocyte transfusions. Am J Pediatr Hematol Oncol6:247-253,1984

Discussion K. Neuman (Washington, D.C.): The neutropenic cancer patient with anorectal infection is a scenario that is obviously encountered with increasing frequency as adjunctive therapy such as bone marrow transplantation becomes more prevalent. The authors have chosen to focus on the 16 children in their series who had positive Pseudomonas cultures and

have applied a different treatment plan for these patients. However, 14 of the 16 patients also grew other organisms as well. Were the patients with Pseudomonas different from 11.5 other children with anorectal inflammation? Does it really matter what the cultures grow? Although most of the deaths were in the early years of the study, 15% of the children

ANORECTALiPERlNEAL

INFECTIONS

IN MALIGNANCY

died with a treatment plan that did not include surgery. Do you see any role for surgical intervention such as the child with a localized collection, extensive gangrene as the patient you just showed, or persistent fistula? Since this experience encompasses 27 years and obviously medical treatment has evolved over that time, I would like to ask your current recommendations for the role in the management plan for rectal examinations in the neutropenic patient, the role of granulocyte macrophage colonies stimulating factor, and granulocyte transfusions. Finally, I would like to second the authors’ overall philosophy of careful and conservative attention to these children and compliment them on their results and presentation. AA. delorimier (San Francisco, CA): I realize that you only looked at the Pseudomonas group, but can you give us a background on other organisms that were involved in this sort of an infection and give us an idea of the mortalities there? I mean, is it the nonoperative treatment in that group of patients as well? I must say we were more aggressive as you were inclined to be in the management of these. I find it disquieting to watch all that necrotic tissue there and just let it slough on itself. I would like to know your feeling about other organisms involved in this sort of infection. D. King (Columbus, OH): I am not at all convinced that your conclusions from reviewing this data are appropriate and that nonoperative management is the way you ought to go. I think you are very fortunate that your patient healed at the end of 5 or 6 months, but I find it hard to believe that that couldn’t have been improved by cutting away some of the diseased tissue and obviously the outcome might not have been any better. The child is alive, but it certainly may not have taken quite so long for him to recover. I find it very hard to believe that taking off dead tissue when there is active infection underneath and promoting drainage wouldn’t be advisable even in these very immunocompromised individuals. A. W. Dibbins (Portland, ME): I used to see a lot of this, but I don’t think I’ve seen a case of this despite seeing a large number of leukemics all treated at our place in the state for a number of years. The only difference that I know is that our pediatric oncologists have absolutely refused to have a pediatric oncology patient ever have a rectal temperature taken. It’s never done anymore, and the disease that you showed and that I saw and feared, certainly when I was in

493

Pittsburgh, has just disappeared. I have literally not seen one in 15 years. 7’.E. Lobe (response): I appreciate all the discussion and comments. Many of you echoed some of my initial concerns on my arrival at St Jude. Let me first answer the questions regarding the other organisms. Of 131 patients, 16 had predominent Pseudomonas infections. The others represent a wide spectrum of expected organisms. Of the 131 patients only one required a colostomy. This was a girl with a severe E coli infection that dissolved her entire perineum and the wall between the vagina and the rectum. She is now about 18 years old and her perineum is unreconstructable. That was the only diversion that we had to perform in 131 patients and that predated our current, more effective antibiotics. Occasionally, you will find a patient, in whom local debridement as Dr King suggested might be indicated, but our concerns are that you might seed this infection, that you might cause some bleeding, and that you might lose control of the situation. The Head of the Infectious Disease Service at St Jude, Dr Walter Hughes, is the fellow who discovered the significance of Pneumocystis, etc. He is the one who held my hands the tightest and kept me from operating on these patients. He has extensive experience with some nasty looking infections and I learned an important lesson. So I really don’t think there is a role for debridement or diversion as long as the patients are improving. Most of these patients have a transient problem. The neutropenia is going to resolve with time and as soon as it does, so will the infection which seems to be fairly superficial. In our experience, none of these are very deep even though there has been much inflammation. As a matter of fact, the child whose slides I showed had edema and erythema above his umbilicus when I first saw him, and I was scared to death just to sit and watch him. Regarding rectal examinations, I too have been taught not to do rectal exams on these children and after seeing some of these infections, either I don’t do them or, under rare circumstances, do them with trepidation. We have used granulocyte stimulating factor and granulocyte transfusion for this problem in selected cases and found these therapys effective. Dr Dibbins, I don’t know why we still see this problem occasionally. I guess we have a lot of cancer patients, maybe more than most institutions, and I guess it is just going to occur every once and a while.