- Email: [email protected]
Evaluation of the risk factors for febrile neutropenia associated with hematological malignancy
J Infect Chemother (2009) 15:174–179 DOI 10.1007/s10156-009-0683-y
© Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2009
ORIGINAL ARTICLE
Yasunori Nakagawa · Kenshi Suzuki · Toru Masaoka
Evaluation of the risk factors for febrile neutropenia associated with hematological malignancy
Received: 23 July, 2008 / Accepted: 27 February, 2009
Abstract Febrile neutropenia (FN) can frequently become a very serious problem. In 2002, Klastersky and colleagues established the Multinational Association for Supportive Care in Cancer (MASCC) score, which consisted of risk factors for conditions that included solid tumors. However, hematopoietic tumors, in comparison to solid tumors, are plagued by such problems as the quantity and quality of abnormalities associated with leukocytes and neutrophils and the requirement for higher dosages of both radio- and chemotherapy. FN is a complication associated with hematological malignancies that can lead to a fatal outcome, but it is avoidable if the appropriate preventive treatment is performed at an early stage. The subjects of the present study consisted of 354 patients with hematopoietic malignancies who were treated at the Japanese Red Cross Medical Center Hospital, Tokyo, between August 2000 and September 2004. They were retrospectively evaluated for the risk factors of FN by applying Wilcoxon’s rank sum test. A scoring index was defined and the patients were classified into high- and low-risk groups before evaluation. The following nine risk factors, which may significantly influence the relationship between the time required for defervescence and the duration of neutropenia – age; hematological diseases; the leukocyte count during the febrile period; the reduction in leukocyte count per day before the onset of FN; the prophylactic administration of antimycotic agents; sterilization of the intestinal tract; and urine albumin content, creatine level, and C-reactive protein (CRP) level – were expressed in points and their sum was termed risk points. The range of risk points was classified as 0–3 and 4– Y. Nakagawa (*) Department of Oncology and Hematology, Japanese Red Cross Medical Center, 4-1-22 Hiroo, Shibuya-ku, Tokyo 150-8935, Japan Tel. +81-3-3400-1311; Fax +81-3-3409-1604 e-mail: [email protected] K. Suzuki Department of Hematology, Japanese Red Cross Medical Center, Tokyo, Japan T. Masaoka Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka, Japan
9. The time required for defervescence was 5.1 days when the risk points were in the range of 0–3 and 8.1 days when the points were in the range of 4–9. These figures were distributed normally and there was a significant difference between the two groups (P = 0.0016). FN associated with hematological malignancies is somewhat different from that related to other malignancies; it is therefore associated with unique risk factors. Most of the risk factors used in the present study can be evaluated objectively. At the onset of FN, they were expressed in points for evaluation. Further prospective studies are needed to determine whether these risk factors are suitable for use in actual cases. Key words Febrile neutropenia · Hematologic malignancy · Risk factor
Introduction Recent progress and developments in chemotherapy and radiotherapy for all types of malignant tumors have resulted in the frequent occurrence of neutropenia with associated fever. In order to address this situation, guidelines concerning febrile neutropenia (FN) were established in the United States in 1990,1 primarily by the Infectious Diseases Society of America. This concept of FN began to be proposed in Japan at the beginning of 1998. Some cases of FN may become serious. In 2002, Klastersky et al.2 designed the Multinational Association for Supportive Care in Cancer (MASCC) score. They identified those patients with a high risk of progressing to a severe infection as the high-risk group and scores were assigned accordingly. In the low-risk group, they noted that there was a less than 5% possibility that these patients would progress to a severe infection.2 However, this system considers risk factors for conditions including solid tumors. Hematopoietic tumors involve not only numerical but also qualitative abnormalities of leukocytes and neutrophils and, in comparison to solid tumors, they may require more intensive chemo- or radiotherapy. Therefore, they cannot be discussed on the same grounds as solid tumors.
175
FN is a complication that may lead to a fatal outcome, but this can probably be avoided if its seriousness is understood, appropriate preventive medication can be given, and treatment is started early. In the present study, the risk factors associated with FN were evaluated retrospectively in patients with malignant hematopoietic diseases who were managed at this institution between August 2000 and September 2004. A scoring index was defined and the patients were classified into highand low-risk groups to verify the classification system.
Others 40(11.3%) AML 131(37.0%) MM 38(10.7%) Total number 354 NHL 55(15.5%) ALL 21(5.9%) MDS 69(19.5%)
Patients and methods (1) FN was defined as a neutrophil count of 1000/μl or less and in a condition in which it may even decrease to below 500/μl; and an axial temperature of more than 37.5°C (oral temperature ≥38°C), while excluding those patients with known causes of fever development, such as those with drug-induced fever, tumor-associated fever, collagen diseases, and allergy.3 (2) A retrospective study was conducted on FN (MASCC low-risk group as controls) in 354 patients with hematopoietic malignant diseases between August 2000 and September 2004. (3) The hematological diseases that formed the background of FN were classified and examined. (4) Wilcoxon’s rank sum test was conducted to evaluate the relationships among the time required for defervescence, duration of neutropenia, and the following factors: age, sex, leukocyte count at onset, neutrophil count, reduction in leukocyte count at onset, the presence of causative agents, disinfection of the digestive tract, prophylactic administration of antimycotic agents (fluconazole [FLCZ] or itraconazole [ITCZ]), the use of granulocyte colony-stimulating factor (G-CSF), and urine albumin content, creatine level, and C-reactive protein (CRP) value. (5) A factor that correlated significantly with the time required for defervescence or the duration of neutropenia was considered to be a risk factor. (6) The patients were scored by risk factors and were then separated into high- and low-risk groups. The time required before defervescence was examined in these groups.
Results Evaluation of the patients’ backgrounds Acute leukemia was found in 42.9% of the patients in the study, and when combined with the number of patients with myelodysplastic syndrome (MDS), the proportion exceeded one half (62.4%). At this institution, the incidence of multiple myeloma was high. Consequently, the number of cases of FN complicated with this condition was relatively large (Fig. 1). The following age distribution was noted: 6.8% of
Fig. 1. Details of the diseases. AML, acute myelogenous leukemia; ALL, acute lymphoid leukemia; MDS, myelodysplastic syndrome; NHL, non-Hodgkin’s lymphoma; MM, multiple myeloma
the patients were 80 years or older; 20.9% were in their 70s; and 24.6% were in their 60s; 52.3% of the patients were age 60 years or older. FN therefore tended to occur more frequently in those aged 60 years or older. FN and risk factors The severity of FN is thought to be associated with prolongation of the neutropenic period or the time required for defervescence. Therefore, those factors that showed a significant correlation with the time required for defervescence or the duration of neutropenia were selected as risk factors. There was no significant difference in the mean time required for defervescence between those aged less than 60 years and those 60 years or older (7.3 days vs 7.9 days; Table 1). When limited to patients with acute myelogenous leukemia (AML) and MDS, the time required for defervescence was 7.2 days for those aged less than 60 years and 9.5 days for those 60 years or older. It was found that the time became prolonged when the patient was 60 years or older, and the condition of FN thus tended to be more serious (P = 0.0455). The risk was independent of sex. The time required for defervescence was 8.9 days when the leukocyte count was less than 100/μl at onset, 8.0 days when it was less than 500/μl, and 7.4 days when it was 500/μl or more (P = 0.0021 and P = 0.0225, respectively). The duration of neutropenia was 10.3 days in the group with the combined use of G-CSF and 14.8 days in the group without; thus showing that G-CSF had a significant effect in reducing the duration of neutropenia (P = 0.0002). The rapidity of progression of leukopenia was considered to be important in the development of FN. Therefore, the reduction in the leukocyte count per day (ΔN) between the day before and the day of onset was computed. The time required for defervescence was 7.3 days in the group with a loss of 500 or more leukocytes/μl and 8.9 days when the leukocyte loss was less than 500/μl or none at all. It was found that the time for defervescence was prolonged in the group that lost leukocytes more rapidly (P = 0.0539).
176 Table 1. Analysis of risk factors stratified by background factors AML + MDS Item to be measured
Age (years)
Classification
Time required before defervescence (days)
Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
60> 60⬉
140 60
7.2 9.5
7.1 9.3
P = 0.0455*
Classification
Time required before defervescence (days)
All patients Item to be measured
Age (years) Leukocyte at onset (μ/l)
Difference in leukocyte from the previous days (μ/l) Cre (mg/dl) Alb (g/dl) CRP (g/dl)
Item to be measured
Cre (mg/dl) Causative agent Prophylactic treatment with antimycotic agents Disinfection of the digestive tract G-CSF
Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
60> 60⬉ 100> 100⬉ 500> 500⬉ −500> −500⬉ 1.5> 1.5⬉ 3.0> 3.0⬉ 3> 3⬉ 10> 10⬉
169 185 42 312 147 207 75 279 328 26 31 323 60 294 249 105
7.3 7.9 8.9 7.5 8.0 7.4 8.9 7.3 7.4 10.5 11.2 7.3 5.3 8.1 7.0 9.0
6.5 7.1 5.9 6.9 5.9 7.4 8.1 6.4 6.7 8.0 8.8 6.5 4.0 7.1 6.6 7.1
P = 0.2851
Classification
Duration of neutropenia (days)
1.5> 1.5⬉ None Yes None Yes None Yes None Yes
P = 0.0021* P = 0.0225* P = 0.0539 P = 0.0412* P = 0.0045** P = 0.0003*** P = 0.0014** Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
328 26 306 48 104 250 119 235 181 173
13.0 8.2 12.5 13.8 13.1 11.4 13.0 11.8 14.8 10.3
12.7 5.2 12.1 14.2 13.7 11.8 13.8 11.6 13.7 10.4
P = 0.0793
It was found that the presence or absence of causative organisms, as detected by blood cultures, was not a risk factor. For prophylactic purposes, when the neutrophil count is below 1000/μl, it is common to carry out disinfection of the digestive tract with polymyxin sulfate B. The duration of neutropenia was 11.8 days in the group that had undergone digestive tract disinfection and 13.0 days in the group that had not. A significant difference was thus noted between the two (P = 0.0045). Itraconazole (ITCZ) or fluconazole (FLCZ) is administered prophylactically when the neutrophil count decreases to less than 1000/μl. The duration of neutropenia was 11.4 days in the group that was treated with these antimycotic
P = 0.6076 P = 0.0011** P = 0.0045** P = 0.0002***
agents, and 13.1 days in the group that did not have this treatment (P = 0.0011). Among the results of laboratory tests, the albumin level, creatine level, and CRP value were found to have some bearing on the time required for defervescence. The time was 7.3 days in patients with an albumin level of 3.0 g/dl or more and 11.2 days when the level was less than 3.0 g/dl. A significant difference was noted between these figures (P = 0.0045). The time required for defervescence was 7.4 days when the creatine level was less than 1.5 mg/dl and 10.5 days when the level was 1.5 mg/dl or more. A significant difference was noted (P = 0.0412). The time required for defervescence was 5.3 days when the CRP value was below 3 g/dl, 7.0 days when the level was below 10 g/dl, and 9.0 days when
177 Table 2. Scoring index obtained by classifying the background factors Age
Point 0 60>
One point 60<
Hematologic diseases Leukocyte count at fever onset Daily fluctuation in WBC count before FN onset Prophylactic treatment with antimycotic agents Sterilization of the digestive tract ALB (g/dl) Cre CRP
Other than MDS or acute leukemia 1000∼500 −500⬉
500∼100 −501⭌
Yes
None
Yes
None
3⬉ 1.5> 0∼3
3> 1.5⬉ 3∼10
Scoring index Score points
Time required before defervescence (days)
0 1 2 3 4 5 6 7 8 9
n
Mean
SD
Min
Median
Max
3 7 14 31 75 95 69 37 16 7
2.0 4.1 6.1 5.2 6.5 7.9 8.5 10.8 7.4 11.4
1.7 2.8 4.2 3.1 4.4 7.1 7.8 10.2 4.9 7.9
1 1 1 1 1 1 1 1 3 3
1 4 5 4 5 5 6 7 5.5 9
4 9 16 13 24 46 46 42 19 27
Time required for defervescence( day )
Table 3. Evaluation by risk factor points
Two points MDS and acute leukemia 100≥
10≥
25 20
( N=354 )
15 10 5 0
0
1
2
3
4
5
6
7
8
9
69
37
16
7
Scoring Index the level was 10 g/dl or more. A significant difference was noted (P = 0.0003, P = 0.0014). Definition and verification of the scoring index The nine risk factors – age, hematological diseases, leukocyte count during the febrile stage, the decrease per day in the leukocyte count prior to the development of FN, prophylactic administration of antimycotic agents, sterilization of the intestinal tract, and urine albumin level, creatine level, and CRP value – expressed in points (Table 2) and their sum as the risk score points, were verified in relation to the time required for defervescence (Table 3; Fig. 2). In expressing the risk factors by points, it was anticipated that the combined use of G-CSF was not likely in many of the patients with acute leukemia or MDS and that FN tends to be more severe in such patients; therefore, 2 points were assigned to these cases. For the leukocyte count during fever and CRP, there were three levels – 0 for no risk and 1 and 2 for increased risk – for the points at which significant differences were noted in relation to an increased severity of FN. For the other items, 0 was assigned to no risk and 1 for risk. In scoring the risk factors, the leukocyte count was used instead of the neutrophil count for the relatively simple and rapid assignment of a score at the onset of FN. When the risk points were grouped from 0 to 3 and 4 to 9, the time required for defervescence was 5.1 days for the former range and 8.1 days for the latter range. A normal
N
3
7
14
31
75
95
Fig. 2. Time required for defervescence at each risk point (all patients)
distribution was recognized; and a significant difference was noted (P = 0.0016; Table 4).
Discussion Through the evaluation of age, one of the risk factors, it was noted that the therapeutic intensity was well adjusted by age for patients with malignant lymphoma; therefore, age did not make a significant difference in the time required for defervescence. Conversely, for patients with acute leukemia and MDS, a potent therapeutic modality is required even at the risk of provoking severe adverse effects, regardless of the age of the patients. Concurrent hematological diseases are important factors in relation to the risk of FN. The severity of therapyassociated neutropenia differs in patients with acute leukemia and those with MDS compared with that in malignant lymphoma and myeloma. In acute leukemia and MDS the extent of neutrophil reduction is more evident. In conditions such as malignant lymphoma and multiple myeloma, where the duration of neutropenia may be reduced by treatment with G-CSF, it was thought that this reduction in the neutropenic period could also reduce the risk that FN will progress to a more severe state. At the same time, the dura-
178 Table 4. Evaluation of the period of fever in relation to high and low risks Subjects
n = 354
Scoring index
Time required before defervescence (days)
Classification
n
Mean
SD
Wilcoxon’s rank sum test
0–1 2–9 0–2 3–9 0–3 4–9 0–4 5–9 0–5 6–9 0–6 7–9 0–7 8–9
10 344 24 330 55 299 130 224 225 129 294 60 331 23
3.5 7.8 5.0 7.8 5.1 8.1 5.9 8.6 6.8 9.2 7.2 9.9 7.6 8.6
2.6 6.8 3.8 6.9 3.4 7.2 4.1 7.8 5.6 8.3 6.2 8.8 6.8 6.1
P = 0.0079** P = 0.0256* P = 0.0016** P = 0.0028** P = 0.0029** P = 0.0120* P = 0.2381
(*: P < 0.05 **: P < 0.01)
tion and dosage of medication, such as antibiotics, could be reduced, which would be an added advantage regarding overall medical costs. On the other hand, in patients with FN associated with acute myelocytic leukemia and MDS, in whom G-CSF may contribute to the exacerbation of the background diseases and cannot therefore be used, the duration of neutropenia is prolonged and can therefore become a risk factor. Although it is conceivable that the dosage of therapeutic agents was already adjusted during chemotherapy conducted in the preliminary stage, the severity of FN has been shown to be proportional to aging. A decline in immunologic potency related to aging is also suspected to have some effects. The effects of latent sex-specific infections, such as prostatitis and cystitis, the latter being common among women, were considered; but it was concluded that sex was not a significant risk factor. The leukocyte count at onset is an important risk factor. It was found that at counts of 500/μl and below 100/μl, FN became more serious; and the leukocyte count at the onset of FN was inversely related to the duration of neutropenia and the severity of this condition. The study also included an observation that the rapidity of leukocyte loss may have had an effect on the status of FN. It was indicated that a dramatic loss of leukocytes each day reflected the severity of FN. According to the observations at this institution, the causative organisms of FN isolated from blood cultures of patients with hematological diseases were gram-positive bacteria in 57.1% and gram-negative bacteria in 42.9%. Causative organisms were found from the blood cultures in approximately 18% of the patients with FN. It was suspected that there may have been problems with the accuracy of the tests in detecting these causative agents. The presence of the causative organisms could not be regarded as a risk factor. The group of patients who had their digestive tracts disinfected was compared with the group who did not. The
disinfection procedure appeared to have contributed to a shortening of the period leading to defervescence. Because an invasion of bacteria through the gastrointestinal mucosa is a serious problem related to the etiology of FN, the disinfection of the digestive system is thus considered to be a significant factor. It was thought that prophylactic treatment with antimycotic agents could contribute to a shortening of the duration of fever. It is said that mycotic infection results in a poor prognosis; in regard to FN, prophylactic treatment with antimycotic agents may very well play an important role in preventing fungal infections. It was shown that an albumin content of less than 3.0 g/dl affected the time required for defervescence in FN. The belief was that the albumin level is an indicator of the nutritious state of a patient and affects the efficacy of antibiotics. A creatine level of 1.5 mg/dl or more seemed to affect the time required for defervescence and the duration of neutropenia. Renal function affects the efficacy or choice of antibiotics and renal dysfunction also contributes to the exacerbation of FN. A CRP value of more than 3, 5, or 10 mg/dl at the onset of FN affected the time required for defervescence as well as the severity of FN. The risk factors that affected the time required for defervescence and the duration of neutropenia were expressed in points to be used as scoring indexes. It was shown that scores from 0 to 3 were representative of patients in the low-risk group, while scores from 4 to 9 indicated those in the high-risk group. In the low-risk group, prospective studies are warranted, and studies should be done to determine whether FN can be controlled by the administration of new quinolone oral preparations.
Conclusion FN associated with hematological malignancies has features that are different from FN associated with malignant solid
179
tumors; therefore, it is necessary to identify these unique risk factors. Most of the risk factors cited in the present study can be objectively evaluated. Furthermore, in order to create an effective scoring index, these risk factors can be expressed in points at the onset of FN. Further prospective studies will be required to confirm whether or not these risk factors are suitable for use in actual cases. Acknowledgments In regard to the evaluation of the risk factors for FN associated with hematological malignancy, we would like to thank Dr. Masaoka, Dr. Kanamaru, Dr. Urabe, and Dr. Tamura of the FN Society for their advice.
References 1. Hughes WT, Armstrong D, Bodey GP, Feld R, Mandell GL, Meyers JD, et al. From the Infectious Diseases Society of America. Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. J Infect Dis 1990;161:381–96.
2. Klastersky J, Paesmans M, Rubenstein EB, Boyer M, Elting L, Feld R, et al. The Multinational Association for Supportive Care in Cancer risk index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 2002;18: 3038–51. 3. Masaoka T. Evidence-based recommendations for antimicrobial use in febrile neutropenia in Japan: executive summary. Clin Infect Dis 2004; 39 (Suppl 1):S49–52.
© Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2009
ORIGINAL ARTICLE
Yasunori Nakagawa · Kenshi Suzuki · Toru Masaoka
Evaluation of the risk factors for febrile neutropenia associated with hematological malignancy
Received: 23 July, 2008 / Accepted: 27 February, 2009
Abstract Febrile neutropenia (FN) can frequently become a very serious problem. In 2002, Klastersky and colleagues established the Multinational Association for Supportive Care in Cancer (MASCC) score, which consisted of risk factors for conditions that included solid tumors. However, hematopoietic tumors, in comparison to solid tumors, are plagued by such problems as the quantity and quality of abnormalities associated with leukocytes and neutrophils and the requirement for higher dosages of both radio- and chemotherapy. FN is a complication associated with hematological malignancies that can lead to a fatal outcome, but it is avoidable if the appropriate preventive treatment is performed at an early stage. The subjects of the present study consisted of 354 patients with hematopoietic malignancies who were treated at the Japanese Red Cross Medical Center Hospital, Tokyo, between August 2000 and September 2004. They were retrospectively evaluated for the risk factors of FN by applying Wilcoxon’s rank sum test. A scoring index was defined and the patients were classified into high- and low-risk groups before evaluation. The following nine risk factors, which may significantly influence the relationship between the time required for defervescence and the duration of neutropenia – age; hematological diseases; the leukocyte count during the febrile period; the reduction in leukocyte count per day before the onset of FN; the prophylactic administration of antimycotic agents; sterilization of the intestinal tract; and urine albumin content, creatine level, and C-reactive protein (CRP) level – were expressed in points and their sum was termed risk points. The range of risk points was classified as 0–3 and 4– Y. Nakagawa (*) Department of Oncology and Hematology, Japanese Red Cross Medical Center, 4-1-22 Hiroo, Shibuya-ku, Tokyo 150-8935, Japan Tel. +81-3-3400-1311; Fax +81-3-3409-1604 e-mail: [email protected] K. Suzuki Department of Hematology, Japanese Red Cross Medical Center, Tokyo, Japan T. Masaoka Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka, Japan
9. The time required for defervescence was 5.1 days when the risk points were in the range of 0–3 and 8.1 days when the points were in the range of 4–9. These figures were distributed normally and there was a significant difference between the two groups (P = 0.0016). FN associated with hematological malignancies is somewhat different from that related to other malignancies; it is therefore associated with unique risk factors. Most of the risk factors used in the present study can be evaluated objectively. At the onset of FN, they were expressed in points for evaluation. Further prospective studies are needed to determine whether these risk factors are suitable for use in actual cases. Key words Febrile neutropenia · Hematologic malignancy · Risk factor
Introduction Recent progress and developments in chemotherapy and radiotherapy for all types of malignant tumors have resulted in the frequent occurrence of neutropenia with associated fever. In order to address this situation, guidelines concerning febrile neutropenia (FN) were established in the United States in 1990,1 primarily by the Infectious Diseases Society of America. This concept of FN began to be proposed in Japan at the beginning of 1998. Some cases of FN may become serious. In 2002, Klastersky et al.2 designed the Multinational Association for Supportive Care in Cancer (MASCC) score. They identified those patients with a high risk of progressing to a severe infection as the high-risk group and scores were assigned accordingly. In the low-risk group, they noted that there was a less than 5% possibility that these patients would progress to a severe infection.2 However, this system considers risk factors for conditions including solid tumors. Hematopoietic tumors involve not only numerical but also qualitative abnormalities of leukocytes and neutrophils and, in comparison to solid tumors, they may require more intensive chemo- or radiotherapy. Therefore, they cannot be discussed on the same grounds as solid tumors.
175
FN is a complication that may lead to a fatal outcome, but this can probably be avoided if its seriousness is understood, appropriate preventive medication can be given, and treatment is started early. In the present study, the risk factors associated with FN were evaluated retrospectively in patients with malignant hematopoietic diseases who were managed at this institution between August 2000 and September 2004. A scoring index was defined and the patients were classified into highand low-risk groups to verify the classification system.
Others 40(11.3%) AML 131(37.0%) MM 38(10.7%) Total number 354 NHL 55(15.5%) ALL 21(5.9%) MDS 69(19.5%)
Patients and methods (1) FN was defined as a neutrophil count of 1000/μl or less and in a condition in which it may even decrease to below 500/μl; and an axial temperature of more than 37.5°C (oral temperature ≥38°C), while excluding those patients with known causes of fever development, such as those with drug-induced fever, tumor-associated fever, collagen diseases, and allergy.3 (2) A retrospective study was conducted on FN (MASCC low-risk group as controls) in 354 patients with hematopoietic malignant diseases between August 2000 and September 2004. (3) The hematological diseases that formed the background of FN were classified and examined. (4) Wilcoxon’s rank sum test was conducted to evaluate the relationships among the time required for defervescence, duration of neutropenia, and the following factors: age, sex, leukocyte count at onset, neutrophil count, reduction in leukocyte count at onset, the presence of causative agents, disinfection of the digestive tract, prophylactic administration of antimycotic agents (fluconazole [FLCZ] or itraconazole [ITCZ]), the use of granulocyte colony-stimulating factor (G-CSF), and urine albumin content, creatine level, and C-reactive protein (CRP) value. (5) A factor that correlated significantly with the time required for defervescence or the duration of neutropenia was considered to be a risk factor. (6) The patients were scored by risk factors and were then separated into high- and low-risk groups. The time required before defervescence was examined in these groups.
Results Evaluation of the patients’ backgrounds Acute leukemia was found in 42.9% of the patients in the study, and when combined with the number of patients with myelodysplastic syndrome (MDS), the proportion exceeded one half (62.4%). At this institution, the incidence of multiple myeloma was high. Consequently, the number of cases of FN complicated with this condition was relatively large (Fig. 1). The following age distribution was noted: 6.8% of
Fig. 1. Details of the diseases. AML, acute myelogenous leukemia; ALL, acute lymphoid leukemia; MDS, myelodysplastic syndrome; NHL, non-Hodgkin’s lymphoma; MM, multiple myeloma
the patients were 80 years or older; 20.9% were in their 70s; and 24.6% were in their 60s; 52.3% of the patients were age 60 years or older. FN therefore tended to occur more frequently in those aged 60 years or older. FN and risk factors The severity of FN is thought to be associated with prolongation of the neutropenic period or the time required for defervescence. Therefore, those factors that showed a significant correlation with the time required for defervescence or the duration of neutropenia were selected as risk factors. There was no significant difference in the mean time required for defervescence between those aged less than 60 years and those 60 years or older (7.3 days vs 7.9 days; Table 1). When limited to patients with acute myelogenous leukemia (AML) and MDS, the time required for defervescence was 7.2 days for those aged less than 60 years and 9.5 days for those 60 years or older. It was found that the time became prolonged when the patient was 60 years or older, and the condition of FN thus tended to be more serious (P = 0.0455). The risk was independent of sex. The time required for defervescence was 8.9 days when the leukocyte count was less than 100/μl at onset, 8.0 days when it was less than 500/μl, and 7.4 days when it was 500/μl or more (P = 0.0021 and P = 0.0225, respectively). The duration of neutropenia was 10.3 days in the group with the combined use of G-CSF and 14.8 days in the group without; thus showing that G-CSF had a significant effect in reducing the duration of neutropenia (P = 0.0002). The rapidity of progression of leukopenia was considered to be important in the development of FN. Therefore, the reduction in the leukocyte count per day (ΔN) between the day before and the day of onset was computed. The time required for defervescence was 7.3 days in the group with a loss of 500 or more leukocytes/μl and 8.9 days when the leukocyte loss was less than 500/μl or none at all. It was found that the time for defervescence was prolonged in the group that lost leukocytes more rapidly (P = 0.0539).
176 Table 1. Analysis of risk factors stratified by background factors AML + MDS Item to be measured
Age (years)
Classification
Time required before defervescence (days)
Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
60> 60⬉
140 60
7.2 9.5
7.1 9.3
P = 0.0455*
Classification
Time required before defervescence (days)
All patients Item to be measured
Age (years) Leukocyte at onset (μ/l)
Difference in leukocyte from the previous days (μ/l) Cre (mg/dl) Alb (g/dl) CRP (g/dl)
Item to be measured
Cre (mg/dl) Causative agent Prophylactic treatment with antimycotic agents Disinfection of the digestive tract G-CSF
Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
60> 60⬉ 100> 100⬉ 500> 500⬉ −500> −500⬉ 1.5> 1.5⬉ 3.0> 3.0⬉ 3> 3⬉ 10> 10⬉
169 185 42 312 147 207 75 279 328 26 31 323 60 294 249 105
7.3 7.9 8.9 7.5 8.0 7.4 8.9 7.3 7.4 10.5 11.2 7.3 5.3 8.1 7.0 9.0
6.5 7.1 5.9 6.9 5.9 7.4 8.1 6.4 6.7 8.0 8.8 6.5 4.0 7.1 6.6 7.1
P = 0.2851
Classification
Duration of neutropenia (days)
1.5> 1.5⬉ None Yes None Yes None Yes None Yes
P = 0.0021* P = 0.0225* P = 0.0539 P = 0.0412* P = 0.0045** P = 0.0003*** P = 0.0014** Test for difference between two independent groups
n
Mean
SD
Wilcoxon’s rank sum test
328 26 306 48 104 250 119 235 181 173
13.0 8.2 12.5 13.8 13.1 11.4 13.0 11.8 14.8 10.3
12.7 5.2 12.1 14.2 13.7 11.8 13.8 11.6 13.7 10.4
P = 0.0793
It was found that the presence or absence of causative organisms, as detected by blood cultures, was not a risk factor. For prophylactic purposes, when the neutrophil count is below 1000/μl, it is common to carry out disinfection of the digestive tract with polymyxin sulfate B. The duration of neutropenia was 11.8 days in the group that had undergone digestive tract disinfection and 13.0 days in the group that had not. A significant difference was thus noted between the two (P = 0.0045). Itraconazole (ITCZ) or fluconazole (FLCZ) is administered prophylactically when the neutrophil count decreases to less than 1000/μl. The duration of neutropenia was 11.4 days in the group that was treated with these antimycotic
P = 0.6076 P = 0.0011** P = 0.0045** P = 0.0002***
agents, and 13.1 days in the group that did not have this treatment (P = 0.0011). Among the results of laboratory tests, the albumin level, creatine level, and CRP value were found to have some bearing on the time required for defervescence. The time was 7.3 days in patients with an albumin level of 3.0 g/dl or more and 11.2 days when the level was less than 3.0 g/dl. A significant difference was noted between these figures (P = 0.0045). The time required for defervescence was 7.4 days when the creatine level was less than 1.5 mg/dl and 10.5 days when the level was 1.5 mg/dl or more. A significant difference was noted (P = 0.0412). The time required for defervescence was 5.3 days when the CRP value was below 3 g/dl, 7.0 days when the level was below 10 g/dl, and 9.0 days when
177 Table 2. Scoring index obtained by classifying the background factors Age
Point 0 60>
One point 60<
Hematologic diseases Leukocyte count at fever onset Daily fluctuation in WBC count before FN onset Prophylactic treatment with antimycotic agents Sterilization of the digestive tract ALB (g/dl) Cre CRP
Other than MDS or acute leukemia 1000∼500 −500⬉
500∼100 −501⭌
Yes
None
Yes
None
3⬉ 1.5> 0∼3
3> 1.5⬉ 3∼10
Scoring index Score points
Time required before defervescence (days)
0 1 2 3 4 5 6 7 8 9
n
Mean
SD
Min
Median
Max
3 7 14 31 75 95 69 37 16 7
2.0 4.1 6.1 5.2 6.5 7.9 8.5 10.8 7.4 11.4
1.7 2.8 4.2 3.1 4.4 7.1 7.8 10.2 4.9 7.9
1 1 1 1 1 1 1 1 3 3
1 4 5 4 5 5 6 7 5.5 9
4 9 16 13 24 46 46 42 19 27
Time required for defervescence( day )
Table 3. Evaluation by risk factor points
Two points MDS and acute leukemia 100≥
10≥
25 20
( N=354 )
15 10 5 0
0
1
2
3
4
5
6
7
8
9
69
37
16
7
Scoring Index the level was 10 g/dl or more. A significant difference was noted (P = 0.0003, P = 0.0014). Definition and verification of the scoring index The nine risk factors – age, hematological diseases, leukocyte count during the febrile stage, the decrease per day in the leukocyte count prior to the development of FN, prophylactic administration of antimycotic agents, sterilization of the intestinal tract, and urine albumin level, creatine level, and CRP value – expressed in points (Table 2) and their sum as the risk score points, were verified in relation to the time required for defervescence (Table 3; Fig. 2). In expressing the risk factors by points, it was anticipated that the combined use of G-CSF was not likely in many of the patients with acute leukemia or MDS and that FN tends to be more severe in such patients; therefore, 2 points were assigned to these cases. For the leukocyte count during fever and CRP, there were three levels – 0 for no risk and 1 and 2 for increased risk – for the points at which significant differences were noted in relation to an increased severity of FN. For the other items, 0 was assigned to no risk and 1 for risk. In scoring the risk factors, the leukocyte count was used instead of the neutrophil count for the relatively simple and rapid assignment of a score at the onset of FN. When the risk points were grouped from 0 to 3 and 4 to 9, the time required for defervescence was 5.1 days for the former range and 8.1 days for the latter range. A normal
N
3
7
14
31
75
95
Fig. 2. Time required for defervescence at each risk point (all patients)
distribution was recognized; and a significant difference was noted (P = 0.0016; Table 4).
Discussion Through the evaluation of age, one of the risk factors, it was noted that the therapeutic intensity was well adjusted by age for patients with malignant lymphoma; therefore, age did not make a significant difference in the time required for defervescence. Conversely, for patients with acute leukemia and MDS, a potent therapeutic modality is required even at the risk of provoking severe adverse effects, regardless of the age of the patients. Concurrent hematological diseases are important factors in relation to the risk of FN. The severity of therapyassociated neutropenia differs in patients with acute leukemia and those with MDS compared with that in malignant lymphoma and myeloma. In acute leukemia and MDS the extent of neutrophil reduction is more evident. In conditions such as malignant lymphoma and multiple myeloma, where the duration of neutropenia may be reduced by treatment with G-CSF, it was thought that this reduction in the neutropenic period could also reduce the risk that FN will progress to a more severe state. At the same time, the dura-
178 Table 4. Evaluation of the period of fever in relation to high and low risks Subjects
n = 354
Scoring index
Time required before defervescence (days)
Classification
n
Mean
SD
Wilcoxon’s rank sum test
0–1 2–9 0–2 3–9 0–3 4–9 0–4 5–9 0–5 6–9 0–6 7–9 0–7 8–9
10 344 24 330 55 299 130 224 225 129 294 60 331 23
3.5 7.8 5.0 7.8 5.1 8.1 5.9 8.6 6.8 9.2 7.2 9.9 7.6 8.6
2.6 6.8 3.8 6.9 3.4 7.2 4.1 7.8 5.6 8.3 6.2 8.8 6.8 6.1
P = 0.0079** P = 0.0256* P = 0.0016** P = 0.0028** P = 0.0029** P = 0.0120* P = 0.2381
(*: P < 0.05 **: P < 0.01)
tion and dosage of medication, such as antibiotics, could be reduced, which would be an added advantage regarding overall medical costs. On the other hand, in patients with FN associated with acute myelocytic leukemia and MDS, in whom G-CSF may contribute to the exacerbation of the background diseases and cannot therefore be used, the duration of neutropenia is prolonged and can therefore become a risk factor. Although it is conceivable that the dosage of therapeutic agents was already adjusted during chemotherapy conducted in the preliminary stage, the severity of FN has been shown to be proportional to aging. A decline in immunologic potency related to aging is also suspected to have some effects. The effects of latent sex-specific infections, such as prostatitis and cystitis, the latter being common among women, were considered; but it was concluded that sex was not a significant risk factor. The leukocyte count at onset is an important risk factor. It was found that at counts of 500/μl and below 100/μl, FN became more serious; and the leukocyte count at the onset of FN was inversely related to the duration of neutropenia and the severity of this condition. The study also included an observation that the rapidity of leukocyte loss may have had an effect on the status of FN. It was indicated that a dramatic loss of leukocytes each day reflected the severity of FN. According to the observations at this institution, the causative organisms of FN isolated from blood cultures of patients with hematological diseases were gram-positive bacteria in 57.1% and gram-negative bacteria in 42.9%. Causative organisms were found from the blood cultures in approximately 18% of the patients with FN. It was suspected that there may have been problems with the accuracy of the tests in detecting these causative agents. The presence of the causative organisms could not be regarded as a risk factor. The group of patients who had their digestive tracts disinfected was compared with the group who did not. The
disinfection procedure appeared to have contributed to a shortening of the period leading to defervescence. Because an invasion of bacteria through the gastrointestinal mucosa is a serious problem related to the etiology of FN, the disinfection of the digestive system is thus considered to be a significant factor. It was thought that prophylactic treatment with antimycotic agents could contribute to a shortening of the duration of fever. It is said that mycotic infection results in a poor prognosis; in regard to FN, prophylactic treatment with antimycotic agents may very well play an important role in preventing fungal infections. It was shown that an albumin content of less than 3.0 g/dl affected the time required for defervescence in FN. The belief was that the albumin level is an indicator of the nutritious state of a patient and affects the efficacy of antibiotics. A creatine level of 1.5 mg/dl or more seemed to affect the time required for defervescence and the duration of neutropenia. Renal function affects the efficacy or choice of antibiotics and renal dysfunction also contributes to the exacerbation of FN. A CRP value of more than 3, 5, or 10 mg/dl at the onset of FN affected the time required for defervescence as well as the severity of FN. The risk factors that affected the time required for defervescence and the duration of neutropenia were expressed in points to be used as scoring indexes. It was shown that scores from 0 to 3 were representative of patients in the low-risk group, while scores from 4 to 9 indicated those in the high-risk group. In the low-risk group, prospective studies are warranted, and studies should be done to determine whether FN can be controlled by the administration of new quinolone oral preparations.
Conclusion FN associated with hematological malignancies has features that are different from FN associated with malignant solid
179
tumors; therefore, it is necessary to identify these unique risk factors. Most of the risk factors cited in the present study can be objectively evaluated. Furthermore, in order to create an effective scoring index, these risk factors can be expressed in points at the onset of FN. Further prospective studies will be required to confirm whether or not these risk factors are suitable for use in actual cases. Acknowledgments In regard to the evaluation of the risk factors for FN associated with hematological malignancy, we would like to thank Dr. Masaoka, Dr. Kanamaru, Dr. Urabe, and Dr. Tamura of the FN Society for their advice.
References 1. Hughes WT, Armstrong D, Bodey GP, Feld R, Mandell GL, Meyers JD, et al. From the Infectious Diseases Society of America. Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. J Infect Dis 1990;161:381–96.
2. Klastersky J, Paesmans M, Rubenstein EB, Boyer M, Elting L, Feld R, et al. The Multinational Association for Supportive Care in Cancer risk index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 2002;18: 3038–51. 3. Masaoka T. Evidence-based recommendations for antimicrobial use in febrile neutropenia in Japan: executive summary. Clin Infect Dis 2004; 39 (Suppl 1):S49–52.