Fungal Colonization and Infection in Children With Acute Leukemia andLymphoma During Induction Therapy

Sevgi Gozdasoglu, MD,* Mehmet Ertem, MD, Zermine Buyukkececi, MD,Sukran Yavuzdemir, MD, Sedef Bengisun, MD, Hatice Ozenci, MD,

Nurdan Tacyildiz, MD, Emel Unal, MD, Gulsan Yavuz, MD, Gulhis Deda, MD, andDerya Aysev, MD

Background. Fungal infection represents agrowing problem in children with hematologicmalignancies. During chemotherapy inducedneutropenia, colonization with fungi is consid-ered a major risk factor for subsequent fungalinfection. The rates and risk factors for mycoticinfections in pediatric oncology patients is un-determined, particularly for centers in develop-ing countries. The aim of this study was toevaluate the rates and risk factors of fungalcolonization in children with acute leukemiaand lymphoma at one of the major pediatrichematology/oncology centers in Turkey. Proce-dure. Fifty-two consecutive children newly di-agnosed with acute leukemia and lymphomaduring intensive remission induction therapywere evaluated for the occurrence of fungalcolonization (defined as at least one positivesurveillance culture) and infection. Results.Thirty-six of the 52 patients (69.2%) were colo-nized by Candida albicans which was the only

fungus isolated from surveillance cultures.There were three (5.8%) proven systemic fungalinfections: two cases of candidemia and onecase of brain abscess with Aspergillus spp. iso-lated from tissue. All patients with fungal colo-nization were receiving prophylactic or cura-tive antibiotics. No significant association wasfound between type of disease and fungal colo-nization, but there was a significant associationwith neutropenia. Conclusions. Our findingssuggest that there is a high rate of fungal colo-nization in children receiving remission induc-tion therapy for acute leukemia and lymphoma.Limiting the use of antibiotics and institutingantifungal chemoprophylaxis may decrease therate, while the early initiation of empiric anti-fungal therapy in patients with fever and sus-pected mycotic colonization may increase sur-vival in these patients. Med. Pediatr. Oncol. 32:344–348, 1999. © 1999 Wiley-Liss, Inc.

Key words: childhood cancer; fungal infection; leukemia; lymphoma;Candida albicans

INTRODUCTION

Fungal infections are a serious and increasingly fre-quent problem in patients with leukemia and lymphoma[1–8]. The rising incidence of fungemia in this popula-tion may be due to the prolonged periods of neutropeniaand mucosal breakdown associated with aggressive che-motherapeutic regimens, the widespread use of broad-spectrum antibiotics, and the use of central venous cath-eters [4–8]. Species of candida have been reported to beresponsible for most of these fungal infections [4,6].Mortality of fungal infections in cancer patients is re-ported to be greater than 50% [9], although the prognosismay be slightly better in the pediatric population [10].

Despite the evidence that fungal infection is becominga major cause of morbidity and mortality, a critical prob-lem limiting effective therapy has been the difficulty inidentifying risk factors associated with its early diagno-sis. Previous studies have suggested that adult patientswith leukemia, with prolonged aplasia and fungal colo-nization during induction therapy, are at high risk ofsubsequent fungal infection [7,11,12]. Despite the evi-dence that fungal colonization predisposes to fungal in-

fection, the rate of colonization in pediatric cancer pa-tients is undetermined, particularly in oncology centers indeveloping countries. The purpose of this study, there-fore, was to investigate the rate and the risk factors re-lated to fungal colonization and infection in pediatricleukemia/lymphoma patients at a major pediatric oncol-ogy center in Turkey.

PATIENTS AND METHODS

Fifty-two consecutive pediatric patients who werenewly diagnosed with acute leukemia or lymphoma atthe pediatric hematology/oncology division of AnkaraUniversity Medical School Children’s Hospital were in-

Pediatric Hematology-Oncology Research Center, Ankara UniversitySchool of Medicine, Dikimevi, Ankara, Turkey

Grant sponsor: Ankara University Research Fund.

*Correspondence to: Sevgi Go¨zdasoglu, Department of Pediatrics, An-kara University School of Medicine, Dikimevi, Ankara, Turkey.

Received 21 May 1997; Accepted 20 November 1998

Medical and Pediatric Oncology 32:344–348 (1999)

© 1999 Wiley-Liss, Inc.

cluded in this study. All patients were treated accordingto the U.S.A. Children’s Cancer Group (CCG) protocolsand were evaluated prospectively during remission in-duction therapy. The patients were aplastic for approxi-mately 1 to 3 weeks during this phase of therapy. Anti-fungal chemoprophylaxis was not used, nor were centralvenous catheters in use at the time of this study in ourcenter. Prophylactic oral ciprofloxacin was given to thepatients when the absolute neutrophil count (ANC) wasless than 500/ml. When the patients developed a fever(oral temperature greater than 38.5°C once, or greaterthan 38.0°C two separate times), they were examined forsigns of localized infection. At this point, cultures of thethroat, urine, blood, and other appropriate sites were ob-tained, and urinalysis and a chest radiograph were per-formed. All patients who developed fever during neutro-penia (ANC < 500/mL) were started on empiric antibiotictreatment (ceftazidime or mezlocilline plus amikacin).When the patients remained neutropenic for more than 7days, with persistent or new febrile episodes, they werereevaluated and intravenous amphoteracin B was begun.Surveillance cultures for fungi consisted of urine, stool,and respiratory specimens (sputum, oropharynx, andmouth) obtained at least 2 weeks after the beginning ofthe induction therapy. These were plated on Sabourauddextrose agar (with gentamicin 100mg/ml). Urine wasstreaked for isolation with a 1:100 calibrated loop topermit quantification, respiratory and stool specimenswere on swabs and were streaked for isolation. Surveil-lance cultures were incubated at 25°C and were exam-ined three times a week for 2 weeks. Gram stains wereused to show the presence of yeast after colony formationwas observed. All of the fungi that were recovered wereidentified using germ tube tests, chlamidaspor presencein corn-meal agar, urease activity, and assimilation andfermentation tests [13].

Fungal colonization was defined as at least one posi-tive surveillance culture. Proven fungal-infection in thisstudy was defined as one or more of the following: 1) theisolation of yeast or filamentous fungi from at least oneblood culture; 2) positive cultures obtained from viabledeep tissue, 3) histopathologic evidence consistent withfungal disease.

RESULTS

Twenty-five of the 52 patients (48.1%) had acute lym-phoblastic leukemia (ALL), 14 (26.9%) had acute myl-eoid leukemia (AML), and 13 (25%) had non-Hodgkinlymphoma (NHL). The ages of the patients ranged from10 to 16 years with both a mean and median of 8.0 years.Thirty-eight patients were male and 14 were female.

During the remission induction therapy of 52 patients,fungal colonization occurred in 36 patients (69.2%). Re-spiratory cultures yielded the highest percentage of posi-

tive cultures among all patients (59.6%); stool cultureswere positive in 15 patients (28.8%), and urine cultureswere positive in only three patients (5.8%). Only candidaspecies all of which wereCandida albicanswere recov-ered from the surveillance cultures.

When fungal colonization rates were examined ac-cording to underlying disease, no statistically significantdifferences were found among the three diagnoses. Six-teen of the 25 (64%) ALL, 11 of the 14 (78.8%) AML,and nine of the 13 (69.2%) NHL patients had fungalcolonization.

Thirty-four of the patients were neutropenic when sur-veillance cultures were obtained. The rate of fungal colo-nization in patients with neutropenia was significantlyhigher; 28 of 34 (82.4%) neutropenic patients had fungalcolonization, while in only eight of the 18 (44.4%) non-neutropenic patients, fungal colonization was found (P <0.05).

All patients had been on antibiotics (prophylactic orempiric treatment for fever and neutropenia) for morethan a week during remission induction therapy. Fourty-seven of these patients were still on antibiotics at the timethe surveillance cultures were obtained.

There were three (5.8%) proven fungal infections in52 patients. Two of these three had positive blood cul-tures for Candida albicans,and they died of systemiccandida infection despite empiric antifungal therapy withamphoteracin B started on day 7 of fever and neutrope-nia. The third patient developed multiple fungal ab-scesses in the brain after a clinical course of pneumonia(Fig. 1). The largest of the abscesses was surgically ex-cised. Direct smear of the content revealed numerousseptate hyphae strongly suggestive of aspergillus as wasconfirmed by later histology and culture of this speci-men. Empiric liposomal amphotericin B had been startedbefore surgery and flucytosine was added to the therapypostoperatively. This patient responded well to the treat-ment and remained clinically well without neurologicsequellae (Fig. 2).

DISCUSSION

Reports from the western world over the last threedecades have indicated that the incidence of fungal in-fection has risen dramatically, with high mortality ratesthat persist, despite advances in antifungal therapy[4,7,10,14,15]. The situation in developing countries isnot well documented, but, in our institution, fungal in-fections are noted as a serious and increasingly commonproblem. In a study that we have recently completed onright atrial catheter-related complications, fungi ac-counted for 17.8% of all organisms isolated from bloodcultures in pediatric oncology patients with right atrialcatheters [16].

Despite the evidence that fungal infection is becoming

Fungal Colonization in Pediatric Oncology 345

a major cause of morbidity and mortality, a critical prob-lem limiting effective therapy has been the difficulty inearly diagnosis. Other investigators have identified riskfactors such as prolonged neutropenia, presence of cen-tral venous catheters, and the use of antibiotics [4,8].Adult patients with acute leukemia and fungal coloniza-tion during induction therapy are at high risk for fungalinfection and might benefit from early, empiric antifun-gal therapy [7,11,12]. In our study, we investigated theoccurrence of fungal colonization and infection in pedi-atric leukemia and lymphoma patients undergoing induc-tion chemotherapy in the light of previously reported riskfactors.

Although fungal colonization has been shown to be animportant risk factor for the development of invasiveinfections, the definition of fungal colonization has dif-fered in most studies. Some studies required at least twopositive surveillance cultures [6,7]. Other investigatorsdefined fungal colonization as one positive surveillanceculture as we did [11,12]. Our colonization rate wasfound to be 69.2% (36/52). When defined as two culturesfrom the same site on 2 separate days, Schwartz et al.found the overall fungal colonization rate to be 58.2% in54 adult patients with AML during induction therapy [7].Wiley et al. analyzed 72 children with acute leukemia

during induction therapy and found an overall fungalcolonization rate (two or more positive surveillance cul-tures) of 47.2% [6]. The differences in the definition offungal colonization and patient population make com-parisons difficult. Nevertheless, in our study as well asthose of others, the fungal colonization rate was duringremission induction therapy for acute leukemia.

Since acute leukemia and lymphoma are the mostcommon underlying malignant diseases in patients withfungal infection [17], only patients with those entitieswere included in our study. Horn et al. reported thatpatients with leukemia were 2.5 times more likely thanlymphoma patients to develop fungal infection [17]. Inour study, when fungal colonization rate was examinedfor each disease, no difference was found between pa-tients with leukemia and lymphoma (69.2 vs. 69.2%).Currently, chemotherapy protocols for patients with lym-phoma are as intensive as those for patients with leuke-mia, which may explain why patients in our study withthese diseases had similar rates of fungal colonization.

Some of the major risk factors for fungal infectionreported in other studies such as the use of central venouscatheters and the administration of parenteral nutritionwere not applicable for our patients. However, other riskfactors such as neutropenia and the use of antibiotics

Fig. 1. MRI of brain showing the large cerebral abscess (arrow).Fig. 2. Brain MRI after antifungal therapy and surgical excision ofabscess.

346 Gozdasoglu et al.

were frequent. All of our patients received prophylacticantibiotics for severe neutropenia or empiric antibiotictreatment for fever and neutropenia. At the time the sur-veillance cultures were taken, 47 of the 52 patients werestill on antibiotics. Therefore, an analysis that wouldcompare the fungal colonization rate in patients receivingantibiotics and those who did not receive antibioticscould not be performed. All of our patients had a neu-tropenic period lasting at least a week during remissioninduction therapy. At the time the surveillance cultureswere taken, 34 of the 52 patients were neutropenic. In ourstudy, the rate of fungal colonization in patients withneutropenia was significantly higher than patients with-out neutropenia.

As others have reported, respiratory cultures in ourpatients yielded the highest percentage of positive cul-tures (59.6%), followed by positive stool cultures in28.8% of the patients [11,12]. Previous research hasdocumentedCandida albicansto be the most commonlyisolated fungal organism. Colonization withCandidatropicalis, Candida parapsilosis, Candida capitatus,andrarely with Aspergillus species have been reported aswell [2–4,7,11,12]. In our sample, prophylactic antifun-gal therapy was not administered, and all colonizationswere withCandida albicans.The widespread us of anti-fungal agents to prevent infections caused byCandidaalbicansin other studies may explain the differences inthese findings. It is also well known thatCandida parap-silosishas a high affinity to central venous catheters usedfor total parenteral nutrition, but these were not in use byus [3,4].

We documented proven systemic fungal infections inthree of the 52 patients evaluated during remission in-duction therapy (5.8%). Two had candidemia, with morethan one positive blood culture, and multiple sites ofcolonization. The third developed brain abscesses withAspergillusspecies isolated from tissue and with nega-tive surveillance cultures. The rate of proven fungal in-fection found in this study is clearly less than that ob-served by other authors who reported 13 to 22% rates ofproven fungal infection in their patients with acute leu-kemia during remission induction therapy [6,7]. This dis-crepancy may be due to the sample selection; we in-cluded patients with lymphoma who may be at lower riskfor fungal infection and excluded patients in relapse whomay be at higher risk. Another possible reason may bethe failure to isolate fungi from blood cultures and theinfrequent use of deep tissue biopsies in our institution.

In the literature, the prognosis of persisting candi-demia in a neutropenic patient has been shown to be verypoor [4,10]. Our patient with candidemia died of infec-tion despite treatment with amphotheracin B. The mor-tality rate for cerebral aspergillosis in highly immuno-compromised patients has been established as close to100% [18]. There are only a few case reports of success-

ful treatment of cerebral aspergillosis so our good resultwith a combination of abscess drainage, liposomal am-photeracin B, and a short course of flucytosine treatmentis of interest [19,20].

In conclusion, our study suggests that the rate of fun-gal colonization is high in children with leukemia andlymphoma during remission induction therapy. Limitingthe use of antibiotics and administering antifungal che-moprophylaxis may reduce the rate of fungal coloniza-tion and infection. Given the high mortality associatedwith fungal infections, we believe empiric therapy withamphoteracin B should be considered early in the eventof fever unresponsive to antibiotics and persistent neu-tropenia. Early initiation of specific antifungal therapymay increase the survival of these patients.

ACKNOWLEDGMENTS

We dedicate this study to the honor of Prof. Dr. AyhanCavdar, the pioneer of pediatric oncology in Turkey.

REFERENCES

1. Berkel I, Say B, Tinaztepe B. Pulmonary aspergillosis in a childwith leukemia. N Engl J Med 1963;269:893–896.

2. Marina NM, Flynn PM, Rivera GK, Hughes WT.Candida tropi-calis andCandida albicansfungemia in children with leukemia.Cancer 1991;68:594–599.

3. Girmenia CG, Martino P, Bernadis FD, et al. Rising incidence ofCandida parapsilosisfungemia in patients with hematologic ma-lignancies: clinical aspects, predisposing factors, and differentialpathogenicity of the causative strains. Clin Infect Dis 1996;23:506–514.

4. Martino P, Girmenia C, Micozzi A, et al. Fungemia in patientswith leukemia. Am J Med Sci 1993;306:225–232.

5. Meunier F, Aoun M, Bitar N. Candidemia in immunocompro-mised patients. Clin Infect Dis 1992;14(suppl 1):120–125.

6. Wiley JM, Smith N, Leventhal BG, et al. Invasive fungal diseasein pediatric acute leukemia patients with fever and neutropeniaduring induction chemotherapy: a multivariate analysis of riskfactors. J Clin Oncol 1990;8:280–286.

7. Schwartz RS, Mackintosh FR, Schrier SL, Greenberg PL. Multi-variate analysis of factors associated with invasive fungal diseaseduring remission induction therapy for acute myelogenous leuke-mia. Cancer 1984;53:411–419.

8. Richet HM, Andremont A, Tancrede C, et al. Risk factors forcandidemia in patients with acute lymphocytic leukemia. Rev In-fect Dis 1991;13:211–215.

9. Fraser VJ, Jones M, Dunkel J, et al. Candidemia in a tertiary carehospital: epidemiology, risk factors, and predictors of mortality.Clin Infect Dis 1992;15:414–421.

10. Stamos JK, Rowley AH. Candidemia in a pediatric population.Clin Infect Dis 1995;20:571–575.

11. Sandford GR, Merz WG, Wingard RJ, et al. The value of fungalsurveillance cultures as predictors of a systemic fungal infection.J Infect Dis 1980;142:503–509.

12. Martino P, Girmenia C, Venditti M, et al. Candida colonizationand systemic infection in neutropenic patients. A retrospectivestudy. Cancer 1989;64:2030–2034.

Fungal Colonization in Pediatric Oncology 347

13. Fromtling RA. Fungi In: Balows A, Hausler WJ Jr, Herrmann KL,Isenberg HD, Shadomy HJ, editors. Manual of clinical microbi-ology (5th ed). Washington, DC: American Society for Microbi-ology; 1991. p 618–625.

14. Bodey GP. Fungal infections complicating acute leukemia. JChronic Dis 1966;19:667–687.

15. Mirsky HS, Cuttner J. Fungal infections in acute leukemia. Cancer1972;30:348–352.

16. Ertem M, Yavuz G, U¨ nal E, et al. Complications of external rightatrial catheters in pediatric cancer patients (abstr). Med PediatrOncol 1997;29:447.

17. Horn R, Wong B, Kiehn TE, Armstrong D. Fungemia in a cancer

hospital: changing frequency, earlier onset, and results of therapy.Rev Infect Dis 1985;7:646–655.

18. Denning DW, Stevens DA. Antifungal and surgical treatment ofinvasive aspergillosis: review of 2121 published cases. Rev InfectDis 1990;12:1147–1201.

19. Coleman JM, Hogg GG, Rosenfeld JV, Waters KD. Invasive cen-tral nervous system aspergillosis: cure with lipozomal amphotera-cin B, itraconazole and radical surgery-case report and review ofliterature. Neurosurgery 1995;36:858–863.

20. Green M, Wald ER, Tzakis A, et al. Aspergillosis of the CNS ina pediatric liver transplant recipient: case report and review. RevInfect Dis 1991;13:653–657.

348 Gozdasoglu et al.