Archives of Medical Research 31 (2000) 388–392
ORIGINAL ARTICLE
Microbiological Findings in Febrile Neutropenia Jesús Gaytán-Martínez,* Eduardo Mateos-García,* Evelia Sánchez-Cortés,** José González-Llaven,** Luis J. Casanova-Cardiel* and José L. Fuentes-Allen* *Departamento de Infectología, Hospital de Infectología, and **Departamento de Hematología, Hospital de Especialidades, Centro Médico Nacional La Raza, Instituto Mexicano del Seguro Social (IMSS), México, D.F., Mexico Received for publication September 7, 1999; accepted March 3, 2000 (99/154).
Background. This study was carried out to assess the isolation rate of bacterial and fungal causative agents in Mexican neutropenic adults with hematological neoplasia. Methods. A prospective observational survey involving 120 consecutive episodes of febrile neutropenia during 1 year was carried out. These episodes were observed in 630 patients discharged with diagnoses of leukemia or lymphoma, or after bone-marrow transplantation. Results. At least one pathogen was isolated in 42 of 120 episodes (35%), and was present in 39 patients with acute myeloid leukemia (AML) (43%), acute lymphoblastic leukemia (ALL) (23%), and in patients who underwent bone-marrow transplantation (20%). Primary bacteremia was the most frequent cause of fever (24 episodes, 57%), followed by intravascular device–related infections (5 episodes, 17%), and soft-tissue infections (5 episodes, 15%). Escherichia coli (33%) was the most frequently isolated agent of primary bacteremia, followed by coagulase-negative Staphylococcus (29%), and Klebsiella oxytoca (16%). Fungal infection was responsible for five events (4%): two episodes of pneumonia (Penicillium marneffei and Aspergillus fumigatus, one event each); two cases of fungemia, one due to Candida tropicalis and one to Rhodotorula gluttinis, and one cryptococcal meningitis event. Conclusions. The isolation rate, approximately 30%, was in accordance with previous reports; similar percentages of Gram-positive and Gram-negative isolates were found. A remarkably low rate of viridans group streptococci and fungal agents was observed, despite the fact that neutropenia is the main risk factor for infection due to these agents. Studies reporting local microbiological findings are necessary because they support an antibiotic choice for prophylaxis or therapy more accurately than reports from other areas. © 2000 IMSS. Published by Elsevier Science Inc. Key Words: Febrile neutropenia, Bacteremia, Bone-marrow transplantation, Leukemia.
Introduction Neutropenia is the strongest risk factor for infection in patients receiving chemotherapy or undergoing bone-marrow transplantation, and fever is the predominant feature of infection (1,2). Forty to 70% of patients with a neutrophil count of ⬍500/mm3 will develop fever and/or infection. Severe infections are common, and bacteremia usually develops when neutrophils fall to ⬍100/mm3 (3,4). Address reprint requests to: José L. Fuentes-Allen, P.O. Box 85-099, 10200 México, D.F., México. Tel.: (⫹525) 659-1606; FAX: (⫹525) 6590306; E-mail:
[email protected]
Identification of specific microorganisms is possible in only 22–39% of cases, probably due to previous empirical antibiotic prophylaxis, which was mandatory in most of these cases (2,3,5). Prophylactic antimicrobial therapy is responsible for the morbidity and mortality decrease in patients undergoing chemotherapy or bone-marrow transplantation; however, this strategy is not without risk, mainly due to antibiotic side effects, and also because of its ecological impact on the environment, specifically the emergence of strains resistant to many antibiotics (6). Twenty years ago, Gram-negative isolates were the most frequently isolated pathogens in neutropenia and fever. Since
0188-4409/00 $–see front matter. Copyright © 2000 IMSS. Published by Elsevier Science Inc. PII S0188-4409(00)00 0 8 0 - 1
Gaytán-Martínez et al. / Archives of Medical Research 31 (2000) 388–392
the 1980s, Gram-positive isolates have become more frequent, presently accounting for 55–66% of bacteremia episodes. Coagulase-negative Staphylococcus and Staphylococcus aureus are the most frequently isolated agents in patients with intravascular-device infections, protracted and aggressive chemotherapy, prophylactic usage of antibiotics, and disruption of gastrointestinal mucosae (1,2,6–8). Infections caused by fungi have also been reported with increasing frequency, and are important contributors to morbidity and mortality in these patients (8–10). When prophylaxis with antibiotics becomes mandatory, the empirical choice of antimicrobial agent should be made on the basis of the microbiology laboratory’s statistical data, collected from previous cases at the same hospital, including susceptibility patterns to antibiotics. Knowledge of the microenvironment ecology would allow for a less empirical and a more rational therapeutic choice (11). Despite the increasing number of patients undergoing chemotherapy and bone-marrow transplantation in Latin American countries, information is lacking about the ecology of infections in neutropenic patients when they become febrile. In Mexico, only two studies concerning infection in immunocompromised patients have been published. The information included in these reports is not sufficient for assessing the clinical features and microbiological findings of our patients. Also, information is scarce in the pediatric population (12–14). The objective of the present study was to characterize the etiological microorganisms causing fever in neutropenic patients in a Mexico City hospital.
Patients and Methods Patient population. Neutropenic patients seen at the Departamento de Hematología, Hospital de Especialidades, Centro Médico Nacional La Raza, a tertiary-care facility in Mexico City, during the year 1997, were enrolled. Patients were admitted to receive chemotherapy for hematological neoplasia, and patients who underwent bone-marrow transplantation were evaluated daily. Patients with fever and neutropenia were included after two blood cultures were taken. Patients with only clinical diagnosis of infection who had no blood culture were not included. Clinical definitions. Neutropenia was defined as a neutrophil count of ⬍1000/mm3, and severe neutropenia when this count was ⬍500/mm3. Fever was defined as a body temperature of ⱖ38.5⬚C. Bacteremia was defined as the presence of clinical symptoms in association with one, of a set of two, positive cultures when the isolated organism was a Gram-negative agent or associated with two or more positive blood cultures when the organism isolated was a Gram-positive agent. Fungemia was considered when the isolate was a fungal or yeast species in one or more blood cultures. Intravenous catheter–related infection was diagnosed when a cath-
389
eter-tip culture was positive with or without pain, skin redness, or pus at the catheter insertion site or adjacent vein. A diagnosis of mucositis was considered when ulceration was present in ⬎25% of the oral surface on a clinical basis. All patients were under prophylactic ciprofloxacin therapy (500 mg b.i.d.) or trimethoprim–sulfamethoxazole (160/800 mg b.i.d.) once neutropenia began. When fever developed, patients were placed under ceftazidime therapy (1 g t.i.d.) plus amikacin (15 mg/kg/qd or adjusted according to creatinine clearance). Specimen handling. Before antibiotics were started, at least three blood cultures were obtained from each patient. When respiratory or urinary symptoms, meningeal signs, or neurological complications arose and soft-tissue infections developed, an ad hoc culture specimen was obtained, that is, sputum or broncoalveolar lavage (BAL) fluid, urine, cerebrospinal fluid obtained by lumbar tap, and tissue biopsy or through syringe aspiration, respectively. The cerebrospinal fluid (CSF) specimen also was processed for cytological and chemical analysis. Culture specimens were processed in an automatic device (BacT Alert System, Organon Teknika Corp., Durham, NC, USA). Mycobacterium sp. and fungi were identified by inoculating the specimens into bottles with Lowenstein– Jensen medium or onto Sabouraud and/or Micosel media plates, respectively. BAL was performed when sputum samples were not useful for diagnosis, in accordance with Murray’s criteria. A preliminary examination of both samples, sputum and BAL fluid was done through one Gram-stained smear and a 10% potassium hydroxide (KOH)-stained smear. As specimens were obtained, they were inoculated onto blood agar, MacConkey agar, and Sabouraud and Lowenstein–Jensen media. Specimens from patients with softtissue infections were inoculated into agar-soya-tripticase with polyanetol sodium sulfonate and activated charcoal. When biopsies were performed, quantitative cultures were done. Bacterial identification was performed by biochemical methods. Susceptibility testing. Testing was performed through the dilution method in agar plates. Statistical analysis. Simple frequencies of fever and neutropenia as cases per 100 discharges were calculated. Isolation rates are expressed as percentages of isolates/febrile neutropenia events. Demographic and microbiological frequency data are expressed as mean ⫾ standard deviation (SD).
Results One hundred twenty consecutive episodes of febrile neutropenia were detected in 630 discharge cases from the Hematology Department (rate 19 episodes per 100 discharges). A positive culture was obtained in 42 of 120 episodes of febrile
390
Gaytán-Martínez et al./ Archives of Medical Research 31 (2000) 388–392
Table 1. Microorganisms isolated in 42 episodes of febrile neutropenia Microorganism
No.
Percentage
Staphylococcus coagulase(-), including Staphylococcus epidermidisa Escherichia coli Klebsiella oxytoca Pseudomonas aeruginosa Enterobacter cloacae Enterococcus faecalis Morganella morganii Hafnia alvei Edwarsiella tarda Klebsiella pneumoniae Bacillus cereus Streptococcus mitis Staphylococcus aureus Corynebacterium renale Enterococcus faecium Rhodotorula gluttinis Cryptococcus neoformans Candida tropicalis Penicillium marneffei Aspergillus fumigatus Total
13 10 4 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 47
28 21 9 5 5 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 100
a
Strains of some Staphylococcus isolates remained unidentified.
neutropenia observed in 39 patients (isolation rate 35%). The mean age of the patients was 33 years (range 16–72 years). There were 24 men (61%) and 15 women (39%). Underlying diseases were acute myeloblastic leukemia (43%), acute lymphoblastic leukemia (23%), and bone-marrow transplantation (20%); other diagnoses, such as lymphoma, aplastic anemia, chronic granulocytic leukemia, and hairy-cell leukemia, accounted for the remaining 13%. Median followup time was 25 ⫾ 10 days and average duration of neutropenia was 10 ⫾ 8 days. Forty-two positive cultures yielded 47 organisms. Bacterial isolates represented 89%, and fungi represented 11%. From bacterial isolates, 48% were Gram-positive organisms, and 52% were Gram-negative bacilli. In 52% of cultureproven episodes (25 events), primary bacteremia was considered the cause of fever, followed by intravascular device infections in 17% (8 events), and by soft-tissue infection in 15% (6 events).
During the study there was neither an outbreak nor a cluster of cases due to a specific microorganism. Isolated agents are shown in Table 1. Predominance of coagulase-negative Staphylococcus and E. coli was evident. Primary bacteremia was the predominant infection, followed by intravascular device– associated infections, and by soft-tissue infections. Patterns of bacterial susceptibility to antibiotics are shown in Tables 2 and 3. Gram-negative isolates exhibited a high rate of resistance to antibiotics. Two of four Klebsiella oxytoca isolates were found resistant to aminoglycosides; two of four resistant to ciprofloxacin; and three of four resistant to ampicillin, carbenicillin, cefotaxime, and ceftazidime. All four isolates of Klebsiella oxytoca were sensitive to imipenem and resistant to trimethoprim–sulfamethoxazole. The only Klebsiella pneumoniae isolate was resistant to trimethoprim–sulfamethoxazole as well. All ten isolates of Escherichia coli showed resistance to ampicillin: nine of ten isolates were resistant to trimethoprim–sulfamethoxazole; four of ten to ciprofloxacin; two of ten to cefotaxime; two of ten to ceftazidime, and one of ten showed resistance to imipenem. Both Pseudomonas aeruginosa isolates showed sensitivity to amikacin, carbenicillin, ceftazidime, ciprofloxacin, and imipenem. Antimicrobial susceptibility testing for Gram-positive isolates showed that all Staphylococcus epidermidis isolates were resistant to ampicillin and trimethoprim–sulfamethoxazole: three of five to cefotaxime; three of five to ciprofloxacin; one of five to dicloxacillin, and one of five to imipenem. All five isolates showed sensitivity to vancomycin. Species identification was unavailable in seven coagulasenegative Staphylococcus isolates. Susceptibility tests showed all these isolates resistant to ampicillin: three of seven showed resistance to amikacin, dicloxacillin, and imipenem; and one of seven to cefotaxime, ciprofloxacin, and trimethoprim–sulfamethoxazole. All seven isolates showed good sensitivity to vancomycin. The sole Staphylococcus aureus isolate was resistant to ampicillin and cefotaxime and showed good sensitivity to dicloxacillin and vancomycin. In regard to Enterococcus strains, one E. faecium and two E. faecalis strains were detected. The E. faecium strain showed resistance to amikacin, cefotaxime, imipenem, and
Table 2. Susceptibility patterns in Gram-negative isolates (percentage of resistance to tested antibiotics) Bacteria
AK
AMP
CRB
CTX
CEFT
CPR
IMP
TMP
Klebsiella oxytoca Escherichia coli Morganella morganii Pseudomonas aeruginosa Enterobacter cloacae Edwarsiella tarda Klebsiella pneumoniae Hafnia alvei
50 10 100 0 50 0 0 0
75 100 100 100 100 100 100 0
75 50 100 0 100 100 100 0
75 20 0 100 100 0 100 0
75 20 0 0 100 0 0 0
50 40 0 0 0 0 0 0
0 10 0 0 0 0 0 0
100 90 100 100 100 0 100 0
AK ⫽ amikacin; AMP ⫽ ampicillin; CRB ⫽ carbenicillin; CTX ⫽ cefotaxime; CEFT ⫽ ceftazidime; CPR ⫽ ciprofloxacin; IMP ⫽ imipenem; TMP ⫽ trimethoprim–sulfamethoxasole.
Gaytán-Martínez et al. / Archives of Medical Research 31 (2000) 388–392
391
Table 3. Susceptibility patterns in Gram-positive isolates (percentage of resistance to tested antibiotics) Bacteria
AK
AMP
CTX
Staphylococcus epidermidis Staphylococcus coagulase(-) Streptococcus mitis Staphylococcus aureus Corynebacterium renale Enterococcus faecium Enterococcus faecalis
0 38 0 0 0 100 50
100 100 0 100 0 0 0
60 12 0 100 100 100 100
CPR
DCX
IMP
TMP
VAN
60 12 0 0 0 0 100
20 38 0 0 100 – –
20 38 0 0 s 100 100
100 12 0 0 0 100 100
0 0 0 0 0 0 0
AK ⫽ amikacin; AMP ⫽ ampicillin; CTX ⫽ cefotaxime; CIPR ⫽ ciprofloxacin; DCX ⫽ dicloxacillin; IMP ⫽ imipenem; TMP ⫽ trimethoprim–sulfamethoxasole; VAN ⫽ vancomycin.
trimethoprim–sulfamethoxazole; in addition, the strain was sensitive to ampicillin, ciprofloxacin, and vancomycin. The same pattern was shown by the two E. faecalis isolated strains. The only Streptococcus mitis isolate was sensitive to ad hoc antibiotics tested.
Discussion This report describes the routine work-up followed in our patients with febrile neutropenia, including bacterial and fungal isolates. For many reasons, mainly the lack of adequate bacterial diagnosis, microbiological data concerning infections in neutropenic patients in developing countries are scarce. The only available data are from two reports, neither specifically related to neutropenia and fever. In an entirely anecdotal study, Paredes-Espinoza reported that either Gram-negative or Gram-positive flora was the main bacterial etiology for infection in immunocompromised patients, describing 26 neutropenic patients, but the author did not identify the frequency of species isolation or other data meaningful to our study. The resulting information is not helpful as a guide for empirical therapy (12). The second report indicates solely, in regard to neutropenic patients, that Gram-positive bacteria were isolated in 75% of patients, and Gram-negative bacteria in 25% of patients. However, this information was obtained from a study that described particularly the use of a type of vascular access in patients with cancer and the authors found neutropenia to be a risk factor associated with intravascular device infection. For this reason, this report is not sufficient with respect to accurate knowledge of hospital ecology and is of little value for a better understanding of infections afflicting neutropenic patients in hospital facilities in Mexico (13,14). As far as we know, this is the first study reporting microbiological findings in febrile neutropenic patients in a tertiarycare facility in Mexico. It is mandatory to obtain basal studies showing local ecological epidemiology, which would be a first step in deciding upon more rational empirical treatments. Following the guidelines recommended by experts from developed countries may potentially be unwise in developing countries, due to the variability among the ecological environ-
ments and different prescription routines for antibiotics and the effects on bacterial selection in a specific nosocomial environment, and also due to the fact that patients in Mexican hospitals differ from patients treated in more developed countries. Observational and descriptive studies are important because they show variation over time and provide clues about the probable etiology of infection in cases of febrile neutropenia. On the basis of reports of this type, local groups will be better prepared to decide on further antimicrobial, combination, or monotherapy. This will be in accordance with a more accurate knowledge of the local environment, and again does not guide therapeutic decisions on expert consensus in this field, which usually only reflects the experiences of other groups. It is necessary to know what is happening worldwide, but it is more important to know what takes place in our patients and to accurately assess our own microbiological findings. In this study, primary bacteremia was found to be the most common infection in febrile neutropenia. It was due mainly to Gram-positive cocci, specifically coagulase-negative Staphylococcus. Interestingly, not one patient with intravenous-device infection had a positive blood culture. There are reports from developed countries showing frequent isolation of Gram-positive bacteria, a differing trend from that found in the past decade, during which Gram-negative bacteria were the predominant genus isolated. With the exception of the viridans group, Staphylococcus isolates are increasing in the developed world (15), but we detected only one isolate from this group. This finding highlights the importance of describing local information, which is in many ways different from data from other centers. The Enterococcus faecalis and E. faecium strains isolated at our center showed good susceptibility to vancomycin, possibly reflecting the low prophylactic usage of this antibiotic, a common routine in the developed world, despite strong recommendations against this practice (16). Also, it is important to note that we selected vancomycin only for patients with a high probability of harboring methicillin-resistant bacteria. These kinds of bacteria are not frequent in our environment, but therapy with vancomycin should be initiated, and dicloxacillin should be deferred un-
392
Gaytán-Martínez et al./ Archives of Medical Research 31 (2000) 388–392
til susceptibility reports are available to avoid the risks associated with undertreating susceptible patients. Unfortunately, many specialists prescribe vancomycin empirically, ignoring the ecological impact this decision will have on the microbial environment. According to the susceptibility patterns exhibited by Klebsiella oxytoca and Escherichia coli to ciprofloxacin (resistance of 50% and 40%, respectively) and to trimethoprim-sulfamethoxazole (resistance of 100% and 90%, respectively), a reconsideration of our policies of prophylaxis with these drugs is mandatory. Such reconsideration has also been recommended by the Fever and Neutropenic Guideline Panel (IDSA) in its latest edition (17). Perhaps by stopping the present prophylactic management policy we can avoid selection of resistant viridans-group streptococci in the future, which is a current problem in the developed world. There are no apparent problems is using the ceftazidime– amikacin combination; only two isolates were susceptible to this combination, which seems to be the reason for the low number of Pseudomonas isolates. Despite the high predisposition to fungal infection induced by neutropenia, only four cases were detected in association with fungi: two cases of pneumonia (one due to Penicillium marneffei and one due to Aspergillus fumigatus), and two cases of fungemia. Surprisingly, neither case of fungemia was due to Candida albicans, the most commonly isolated organism in neutropenic patients. We are presently searching for these fungal infections, due to the high risk factor for neutropenia (18–20). The usefulness of this report as it pertains to our region should be emphasized and, although useful for purposes of comparison, our clinical experience is not necessarily comparable to other centers. Conversely, this also applies to studies coming from abroad. If we do not describe what happens in our own environment, then we cannot make general observations or detect local trends. This study is an invitation for future investigations.
References 1. Donowitz GR. Fever in the compromised host. Infect Dis Clin N Am 1996;10:129. 2. Pizzo PA. Empirical therapy and prevention of infection in the immunocompromised host. In: Mandell GL, Bennett JE, Dolin R, editors. Principles and practice of infectious diseases. New York: ChurchillLivingstone;1995. p. 2686.
3. Da Lalla F. Antibiotic treatment of febrile episodes in neutropenic cancer patients. Drugs 1997;53:789. 4. Hughes WT, Armstrong D, Bodey GP, et al. Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. J Infect Dis 1990;161:381. 5. Pizzo PA. Management of fever in patients with cancer and treatmentinduced neutropenia. N Engl J Med 1993;328:1323. 6. Oreke CH, Dudley MN. -lactam/-lactamase inhibitor combinations: pharmacodynamic considerations and possible role in the management of bacterial infections in the neutropenic host. J Antimicrob Chemother 1998;41(Suppl D):43. 7. Oppenheim BA. The changing pattern of infection in neutropenic patients. J Antimicrob Chemother 1998;41(Suppl D):7. 8. Klastersky J. Science and pragmatism in the treatment and prevention of neutropenic infection. J Antimicrob Chemother 1998;41(Suppl D):13. 9. Denning DW, Evans EGV, Kibbler CC, et al. Guidelines for the investigation of invasive fungal infections in haematological malignancy and solid organ transplantation. Eur J Clin Microbiol Infect Dis 1997;16:424. 10. Lortholary O, Dupont B. Antifungal prophylaxis during neutropenia and immunodeficiency. Clin Microbiol Rev 1997;10:477. 11. Viscoli C. The evolution of the empirical management of fever and neutropenia in cancer patients. J Antimicrob Chemother 1998; 41 (Suppl D):65. 12. Paredes-Espinoza M. Infections in immunocompromised patients in Mexico. Rev Infect Dis 1989;11(Suppl 7):S1644. 13. Volkow P, Sánchez-Mejorada G, Lazo-de-la-Vega S, Vázquez C, et al. Experience of an intravenous therapy team at the Instituto Nacional de Cancerología (Mexico) with a long-lasting, low-cost silastic venous catheter. Clin Infect Dis 1994;18:719. 14. Miranda-Novales MG, Belmont-Martínez L, Villasis-Keever MA, Penagos-Paniagua M, Bernáldez-Ríos R, Solórzano-Santos F. Empirical antimicrobial therapy in pediatric patients with neutropenia and fever. Risk factors for treatment failure. Arch Med Res 1998;29:331. 15. Carratalá J, Alcaide F, Fernández-Sevilla A, Corbella X, Liñares J, Gudiol F. Bacteremia due to viridans streptococci that are highly resistant to penicillin: increase among neutropenic patients with cancer. Clin Infect Dis 1995;20:1169. 16. EORTC and National Cancer Institute of Canada. Vancomycin added to empirical combination antibiotic therapy for fever in granulocytopenic cancer patients. J Infect Dis 1991;163:951. 17. Hughes WT, Armstrong D, Bodey GP, et al. 1997 guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. Clin Infect Dis 1997;25:551. 18. Warnock DW. Fungal infections in neutropenia: current problems and chemotherapeutic control. J Antimicrob Chemother 1998;41(Suppl D):96. 19. Working Party of the British Society for Antimicrobial Chemotherapy. Therapy of deep fungal infection in haematological malignancy. J Antimicrob Chemother 1997;40:779. 20. Gerson SL, Talbot GH, Hurwitz S, Strom BL, Lusk EJ, Cassileth PA. Prolonged granulocytopenia: the major risk factor for invasive pulmonary aspergillosis in patients with acute leukemia. Ann Intern Med 1984;100:345.
