Assessment of Procalcitonin as a Diagnostic Marker of Underlying Infection in Patients with Febrile Neutropenia

Transcription

1 MAJOR ARTICLE Assessment of Procalcitonin as a Diagnostic Marker of Underlying Infection in Patients with Febrile Neutropenia Evangelos J. Giamarellos-Bourboulis, 1,2 Paraskevi Grecka, 1,2 Garyfallia Poulakou, 1 Konstantinos Anargyrou, 1 Nikolaos Katsilambros, 2 and Helen Giamarellou 1 1 4th Department of Internal Medicine and 2 1st Department of Propedeutic Medicine, Athens Medical School, Greece The novel inflammatory marker procalcitonin (PCT) was assessed as an index of infection in patients with febrile neutropenia. Blood samples were obtained from 115 patients with febrile neutropenia for determination of PCT levels before onset of fever and daily until the resolution of fever. The median PCT level on the first day of fever was 8.23 ng/ml in patients with bacteremia, compared with 0.86 ng/ml in patients with localized bacterial infections ( P p.017). The median PCT level on the first day of fever was 2.62 ng/ml in patients with severe sepsis, compared with 0.57 ng/ml in patients with clinically localized infections ( P!.001). A dramatic decrease in PCT levels was documented after resolution of the infection; PCT levels were elevated when the infection worsened. Pronounced PCT levels were also found in patients with fever of unknown origin who were responding to antimicrobial chemotherapy, compared with those not responding to treatment with antibiotics. PCT levels were particularly elevated in patients with bacteremia and severe sepsis. These findings provide new insight into the application of PCT in clinical trials as a diagnostic tool of the severity of an infection in patients with febrile neutropenia and of the need to change antimicrobial regimen. Neutropenia in patients who have undergone antineoplastic chemotherapy is a state of immunosuppression that in extreme vulnerability of the host to numerous microorganisms, which can cause lethal infections. It is well known that febrile neutropenia may be attributed to an underlying infection, to the underlying disorder, or to the administration of drugs or blood products. In many patients, signs of infection are not present. Another problem is the recurrence of fever at the end of a successful course of antimicrobial treatment for a previous infection. To help resolve these questions, various serological markers have been ap- Received 24 February 2000; revised 4 October 2000; electronically published 21 May Reprints or correspondence: Dr. H. Giamarellou, Dept. of Internal Medicine, Sismanoglion General Hospital, 1 Sismanogliou St., Maroussi Attikis, Greece. Clinical Infectious Diseases 2001; 32: by the Infectious Diseases Society of America. All rights reserved /2001/ $03.00 plied, especially C-reactive protein and IL-6 [1]. However, the major disadvantage of their application is their lack of specificity as a result of the effect of the underlying disorder in their production; their high cost is another disadvantage. Procalcitonin (PCT) is a novel peptide that consists of 116 amino acids with the same sequence of the prohormone of calcitonin synthesized in the C cells of the thyroid gland [2]. It is significantly elevated in the serum of patients with bacterial meningitis or with sepsis, but the site of its production remains unclear [3 5]. Because the majority of studies of PCT involve immunocompetent hosts, we have attempted to describe PCT production in patients with febrile neutropenia, particularly in cases of systemic infections comprising bacteremia and severe sepsis. PATIENTS AND METHODS From January 1997 through January 1999, we enrolled a total of 115 hospitalized patients in a prospective 1718 CID 2001:32 (15 June) Giamarellos-Bourboulis et al.

2 Table 1. Clinical characteristics of the 3 groups of patients with febrile neutropenia who were enrolled in the study. Characteristic Microbiologically documented infections (n p 42) Clinically documented infections (n p 34) Fever of unknown origin (n p 39) Epidemiological data Sex Male 32 (76.2) 23 (67.6) 25 (64.1) Female 10 (23.8) 11 (32.4) 14 ( 35.9) Age, mean y SD Underlying malignancy Acute myelogenous leukemia 26 (61.9) 14 (41.1) 18 (46.2) Non-Hodgkin s lymphoma 8 (19.0) 4 (11.8) 3 (7.7) Myelodysplastic syndrome 2 (4.8) 6 (17.6) 3 (7.7) Solid tumor 3 (7.1) 4 (11.8) 4 (10.3) Other 3 (7.1) 6 (17.6) 11 (28.1) Clinical data Granulocyte count at onset of fever, mean SD Type of infection Bacteremia a 28 (63.2) Severe sepsis 9 (21.4) 11 (32.4) Lower respiratory tract infection b 1 (2.4) 19 (55.9) Urinary tract infection c 5 (11.9) 1 (2.9) Soft tissue infection d 7 (16.7) 3 (8.8) Acute tonsillitis e 1 (2.4) 2 (5.9) Acute sinusitis 3 (8.8) Systemic fungosis f 3 (7.1) Successful response to antimicrobial agents 24 (57.1) 22 (57.9) 26 (66.7) Successful response to antifungal agents 6 (14.3) 3 (8.8) 7 (17.9) Deaths 12 (28.6) 10 (26.3) 6 (15.4) NOTE. Data are no. () of patients, unless otherwise indicated. a Escherichia coli in 8 patients, Staphylococcus aureus in 7, Pseudomonas aeruginosa in 3, Enterobacter cloacae in 3, Enterococcus faecalis in 3, and other in 4 patients. b Haemophilus influenzae in 1 patient. c E. coli in 2 patients and E. cloacae in 3. d S. aureus in 4 patients, P. aeruginosa in 2, and E. coli in 1. e Streptococcus pyogenes in 1 patient. f Candida albicans in 2 patients and Aspergillus fumigatus in 1. study. All patients had various hematologic malignancies or solid tumors, and all patients underwent anineoplastic chemotherapy. All patients developed neutropenia (polymorphonuclear neutrophils,!500 cells/mm 3 ) after chemotherapy; they also developed fever. Fever in these patients was defined as any of 3 or more spikes of temperature of 38 C during a 24-h period or a single spike of 38.5 C [5]. A complete clinical evaluation was performed for each patient, which included physical findings; biochemical parameters; cultures of blood, urine, and sputum samples and of specimens of skin lesions; chest and sinus radiographs; and CT scans of the chest and abdomen whenever it was considered necessary. On the basis of the clinical and laboratory findings, patients were divided into the following 3 categories: those with a microbiologically documented infection; those with a clinically documented infection whenever physical or radiological findings were compatible with an underlying infection, but without any microbiological proof; and those with fever of unknown origin (FUO) whenever fever lasted 3 days and no cause of fever could be detected, despite the use of the aforementioned Procalcitonin in Diagnosis of Febrile Neutropenia CID 2001:32 (15 June) 1719

3 Table 2. Daily follow-up of procalcitonin levels in 39 patients with neutropenia who presented with bacteremia or with a localized bacterial infection (excluding patients with systemic fungosis). Procalcitonin levels, ng/ml Time bacteremia localized bacterial infection Before chemotherapy 0.16 (28) ND 0.80 ND (11) ND 0.69 Afebrile on neutropenia 0.05 (28) ND (11) ND 0.28 Days of fever (28) a,b (11) c ND (24) a,b ND (8) c ND (18) a,b ND (8) c ND (14) b,d ND (6) e ND 1.95 NS Resolution of fever 0.37 (20) ND (11) ND 1.0 NOTE. ND, nondetectable; NS, not significant. a P!.001 compared with the values of the same patients in the status of neutropenia before the onset of fever. b P p NS compared with patients with severe sepsis. c P p NS compared with patients with clinical localized infections. d P!.01 compared with the values of the same patients in the status of neutropenia before the onset of fever. e P!.05 compared with the values of the same patients in the status of neutropenia before the onset of fever. tests [6]. Categorization of patients was made without knowledge of PCT levels. For all patients, severe sepsis was defined as sepsis that was associated with organ dysfunction, hypoperfusion, or hypotension. Perfusion abnormalities may include, but were not limited to, lactic acidosis, oliguria, or an acute alteration in mental status [7 9]. Blood samples were obtained from all patients before initiation of chemotherapy; at 48-h intervals, when patients were Table 3. Daily follow-up of procalcitonin levels in 34 patients with neutropenia who presented with severe sepsis or with a localized clinical infection. Time severe sepsis afebrile and neutropenic; and at 24-h intervals, after the patient experienced fever until the fever resolved (temperature, 37.6 C [5]). One part of each blood sample was cultured under both aerobic and anaerobic conditions for 7 days (BAC- TEC; Becton Dickinson), and the other part was centrifuged at 12,000g and 4 C. Serum samples were kept refrigerated at 70 C until they were processed. PCT levels were determined by use of an assay on the basis of immunochemiluminescence Procalcitonin level, ng/ml clinical localized infections Before chemotherapy ND (11) ND (23) ND 1.07 Afebrile on neutropenia 0.35 (11) ND (23) ND 0.80 Days of fever (11) a (23) b ND 16.17! (10) a (21) c ND (8) b (19) c ND (8) b (13) d ND Resolution of fever 0.25 (5) ND (22) ND 1.44 P P NOTE. ND, nondetectable; NS, not significant. a P!.001 b P!.01 c P!.05 d compared with the values of the same patients in the status of neutropenia before the onset of fever. compared with the values of the same patients in the status of neutropenia before the onset of fever. compared with the values of the same patients in the status of neutropenia before the onset of fever. P p NS compared with the values of the same patients in the status of neutropenia before the onset of fever CID 2001:32 (15 June) Giamarellos-Bourboulis et al.

4 Figure 1. Distribution of values of procalcitonin (PCT) on the first day of fever among patients who presented with bacteremia, a localized bacterial infection, severe sepsis, or a localized clinical infection. (BRAHMS Diagnostica; lower detection limit, 0.08 ng/ml) after incubation of the sample for 2 h at 30 C into tubes coated with 2 different monoclonal antibodies. Luminescence was then measured automatically by use of a Berilux Analyzer 250 (Boehring Diagnostics). Concentrations of PCT of!0.5 ng/ml Table 4. Daily follow-up of procalcitonin levels of 39 patients neutropenia who presented with fever of unknown origin in relation to response to antimicrobial chemotherapy. Time positive response to antimicrobial chemotherapy were considered to be negative, according to other authors [3]. All measurements were performed twice. Because values did not follow a Gaussian curve, PCT values were reported as the median values on each day of fever. Values of each febrile patient were compared with those of the same patient while he or she was experiencing neutropenia and before he or she experienced fever by use of the Wilcoxon rank test ( P!.05). microbiologically or clinically documented infections were divided into subgroups of patients with systemic (i.e., bacteremia, severe sepsis, or both) or localized infections (i.e., infections confined to a sole organ), and comparisons between them were performed by use of the Mann- Whitney rank sum test ( P!.05). The changes in PCT levels for each group of patients on consecutive days were evaluated by use of the Friedman s test [10]. RESULTS The epidemiological and clinical data of patients enrolled in the study are shown in table 1. Before the initiation of chemotherapy, the median PCT level was 0.29 ng/ml (range, ng/ml); when the patient was in a neutropenic state before fever began, it was 0.18 ng/ml (range, ng/ml). PCT values of the first 4 days of microbiologically proven infections are shown in table 2. Patients were subdivided into those with bacteremia and those with a localized bacterial infection; we considered the start of the infection to be the first day on which cultures of biologic samples yielded positive. A respective analysis of PCT values of patients with neutropenia who had severe sepsis or a clinically localized infection Procalcitonin levels, ng/ml out response to antimicrobial chemotherapy Before chemotherapy 0.11 (25) ND (14) ND 0.75 Afebrile on neutropenia 0.20 (25) ND (14) ND 0.64 Days of fever (25) a ND (14) b ND (19) a ND (9) a ND 3.62 NS (11) c ND (8) a ND NS (9) a ND (5) b ND 6.48 NS Resolution of fever 0.37 (25) ND (14) ND 3.35 P NOTE. ND, nondetectable; NS, not significant. a P! 0.05 compared with the values of the same patients in the status of neutropenia before the onset of fever. b P p NS compared with the values of the same patients in the status of neutropenia before the onset of fever. c P! compared with the values of the same patients in the status of neutropenia before the onset of fever. Procalcitonin in Diagnosis of Febrile Neutropenia CID 2001:32 (15 June) 1721

5 Figure 2. Follow-up of procalcitonin (PCT) over time in patients with bacteremia and with severe sepsis in relation to the response to the administration of antimicrobial chemotherapy. is presented in table 3; we considered the start of the infection to be the first day on which clinical signs of an infection occurred. The distribution of the values of PCT in patients with a documented infection is shown in figure 1, and follow-up of patients with a systemic infection correlated to their response to antimicrobial chemotherapy is shown in figure 2. PCT values after the first day of the occurrence of FUO in relation to response to antimicrobial chemotherapy are given in table 4. For the group of patients who presented with FUO, fever resolved in 25 patients (64.1) after the administration of an antimicrobial regimen, and PCT values of 10.5 ng/ml were found in 15 patients (60). No response of fever was observed in any of the remaining 14 patients (35.9); of these 14 patients, the PCT level was found to be elevated (10.5 ng/ ml) in 1 (6.7; P p.043). For the group of patients who presented with bacteremia, the median PCT value on the first day of fever in patients with infections caused by gram-negative isolates was ng/ml (range, ng/ml); for patients with infections caused by gram-positive cocci, the median PCT value was 1.29 ng/ml (range, ; P p NS). Nine patients with bacteremia also developed signs of severe sepsis; the median PCT value was ng/ml (range, ng/ml) on the first day of fever ( P p NS, compared with patients without severe sepsis and without bacteremia). Three patients presented with systemic fungosis, 1 presented Table 5. Definitions of sensitivity, specificity, and positive predictive values of different concentrations of procalcitonin determined on the first day of febrile neutropenia for the diagnosis of bacteremia. Procalcitonin, ng/ml No. of true positive bacteremia No. of false negative localized infections No. of false positive No. of true negative Sensitivity, Specificity, PPV, NOTE. PPV, positive predictive value CID 2001:32 (15 June) Giamarellos-Bourboulis et al.

6 Table 6. Definitions of sensitivity, specificity, and positive predictive values of different concentrations of procalcitonin determined on the first day of febrile neutropenia for the diagnosis of severe sepsis. Procalcitonin, ng/ml No. of true positive severe sepsis No. of false negative localized infections No. of false positive No. of true negative Sensitivity, Specificity, PPV, NOTE. PPV, positive predictive value. with pulmonary aspergillosis, and 2 presented with hepatosplenic candidiasis (table 1). Diagnosis of pulmonary aspergillosis was made after the patient underwent bronchoalveolar lavage; the other 2 patients underwent splenectomy before they received their diagnosis. The patient with pulmonary aspergillosis presented with PCT values with a range of ng/ ml (median, 1.26 ng/ml) and had a total duration of fever of 24 days. For the other 2 patients, the range of PCT values was ng/ml (median, 0.26 ng/ml) and ng/ml (median, 0.18 ng/ml), respectively; the total duration of fever was 78 and 18 days, respectively. We have attempted to define the concentration of PCT that might be considered to be a breakpoint to differentiate between bacteremia or severe sepsis and localized infection; the sensitivity, the specificity, and the positive predictive values of these concentrations are presented in tables 5 and 6. To achieve that determination, patients with bacteremia or severe sepsis and elevated PCT values were considered to have a true-positive finding, and those with localized infections without elevated PCT levels were considered to have a true-negative finding. DISCUSSION PCT is a 116 amino acid peptide with the same sequence of the prohormone of calcitonin synthesized by the C cells of the thyroid gland. It is particularly elevated in the serum samples of immunocompetent patients with sepsis and bacterial meningitis [3, 4]; it is also elevated in healthy volunteers after they receive an injection of endotoxin [11], and it is considered to be a specific infective parameter. Although PCT is produced to some extent by WBC [12], which, therefore, renders its diagnostic value in patients with febrile neutropenia doubtful, preliminary by Al Nawas and Shah [2] showed an elevation of PCT levels in 25 patients with neutropenia and sepsis. The present study attempted to describe the kinetics of PCT in patients with febrile neutropenia in order to define its value as a marker of infection in patients with immunosuppression. Before the onset of fever, PCT values remained within normal detection limits both in patients without neutropenia who had a malignancy and in afebrile hosts with neutropenia (median values, 0.29 ng/ml and 0.18 ng/ml, respectively). However, the value of PCT as an index of infection was mainly documented on the first day of the presentation of the infection. The median PCT value on the first day of bacteremia was 8.23 ng/ ml highly statistically significant compared with the median PCT value of 0.86 ng/ml on the first day of a localized bacterial infection (table 2). On the other hand, analysis of PCT failed to detect the 3 patients with systemic fungosis for the following 2 reasons: (1) in 1 patient, not a single PCT value surpassed the threshold of 0.5 ng/ml, and (2) a wide range of values was observed throughout the course of their infection. Transient elevations of PCT levels in patients with systemic fungosis have been reported in 2 cases of disseminated aspergillosis [13]. For patients with neutropenia who had clinically documented infection (table 3), concerning the observed differences between systemic and localized infections are in complete agreement with those documented in patients with a microbiologically proven infection. More precisely, the median PCT level on the first day of severe sepsis was 2.62 ng/ml, a finding that is statistically superior to findings regarding patients with localized infections, for whom median PCT value was 0.57 ng/ml. Of importance, in all patients who presented with either systemic or localized infections (tables 2 and 3), levels of PCT were determined and then compared with the patients status of neutropenia before the onset of the infection. On the basis of the above, it might be hypothesized that, because elevated PCT levels were found on the first day in patients with FUO who responded to antimicrobial chemotherapy (compared with patients who did not respond to the administration of an antimicrobial regimen), elevation of PCT levels was indicative of patients with a probable bacterial infection (table 4). That is further supported by the finding that 60 of the patients with FUO who responded to the administration of a regimen of antimicrobial agents presented with Procalcitonin in Diagnosis of Febrile Neutropenia CID 2001:32 (15 June) 1723

7 elevated PCT values (i.e., 10.5 ng/ml), whereas only 6.7 of those who did not respond to the administration of antimicrobial agents had elevated PCT values. The main problem with the determination of PCT values is the wide variation of the obtained values (figure 1), which do not follow a Gaussian curve model, and which, therefore, make it difficult to establish a threshold with an adequate sensitivity and specificity to indicate the diagnosis of a certain syndrome (tables 5 and 6). Ideally, the applied diagnostic test should posses the maximum available sensitivity, specificity, and positive predictive value [14]. Although there is no one concentration of PCT that meets all 3 criteria, and considering that a sensitivity of 73 has been proposed by other authors as satisfactory for PCT [15], it might be assumed that concentrations of 11.0 ng/ml with 78.6 sensitivity, accompanied by relatively satisfactory criteria for specificity and for positive predictive value, might be helpful for the differential diagnosis of bacteremia and localized bacterial infection. PCT values of 12.0 ng/ml seem adequate for the differentiation between severe sepsis and localized infections, with 90.9 sensitivity, 80.9 specificity, and 76.9 positive predictive value. As the applied concentrations of PCT increase, sensitivity decreases, but specificity and positive predictive value increase; this leads us to assert that concentrations of 12.5 ng/ml in a single patient should raise the suspicion of severe sepsis. However, careful inspection is needed to draw safe conclusions regarding the kinetics of PCT in plasma over time and regarding the influence exerted by the antimicrobial regimens. It has been proposed by other authors that values of PCT remain elevated for 3 days, then start to decrease [2]. However, when considering PCT values in patients with a systemic infection in relation to the therapeutic response (figure 2), it becomes clear that the resolution of the infection is accompanied by a dramatic decrease in PCT levels from the first 24 h, whereas persistence of the infection is accompanied by highly elevated PCT levels in plasma. In patients with FUO who responded to the empirical administration of an antimicrobial regimen, after the second day of therapy, response to treatment was accompanied by a drop of elevated PCT values to values that did not differ from control values (table 4). The we present are in general accordance with those of other authors who have studied immunocompetent hosts to reveal an elevation of PCT values in patients with bacteremia and sepsis [2 5]. To our knowledge, only 2 studies have been published on the importance of PCT as a diagnostic marker in patients with febrile neutropenia. The first study [16] comprises descriptions of 28 cases of neutropenia and 7 cases of bacteremia; this study found elevated PCT concentrations in the patients, compared with patients with FUO, and the authors found that gram-negative bacteremia was accompanied by greater values of PCT, compared with gram-positive bacteremia. The second study [15] involved 33 cases of bacteremia and revealed an enormous elevation of PCT values without any observed differences between gram-positive and gram-negative bacteremias; these findings are in complete agreement with ours. However, the latter study did not involve any patients with sepsis or severe sepsis. In conclusion, the present study revealed that PCT might be a useful diagnostic tool for the early detection of a systemic infection in patients with febrile neutropenia. Values of ng/ml might suggest the existence of a localized infection; values of 11.0 ng/ml might suggest a probable bacterial bloodstream seeding; and values 12.0 ng/ml suggest a septic response. There is no doubt that the presence of febrile neutropenia demands the early start of antimicrobial chemotherapy. However, the we present might support the need to conduct studies that would clarify whether patients with febrile neutropenia accompanied by elevated PCT levels should be managed with an enhanced regimen of antimicrobial agents or whether the regimen that had been administered should be changed according to sustained PCT levels, especially in patients with FUO. References 1. Hansson LO, Lindquist L. C-reactive protein: its role in the diagnosis and follow-up of infectious diseases. Curr Opin Infect Dis 1997; 10: Al Nawas B, Shah PM. Procalcitonin in patients with and without immunosuppression and sepsis. Infection 1996; 24: Assicot M, Gendrel D, Carsin H, Raymond J, Guilbaud J, Bouhon C. High serum procalcitonin concentrations in patients with sepsis and infection. Lancet 1993; 341: Gendrel D, Raymond J, Assicot M, et al. Procalcitonine, protéine C- réactive et interleukine 6 dans les méningites bactériennes et virales de l infant. Presse Med 1998; 27: Whang KT, Steinwald PM, White JC, et al. Serum calcitonin precursors in sepsis and systemic inflammation. J Clin Endocrinol Metab 1998; 83: Gelfand JA, Dinarello CA. Fever of unknown origin: definition and classification. In: Fauci AS, Braunwald E, Isselbacher KJ, et al., eds. Harrison s principles of internal medicine. 14th ed. New York: McGraw Hill 1998: Balk RA. Severe sepsis and septic shock: definitions, epidemiology, and clinical manifestations. Crit Care Clin 2000; 16: Pizzo PA. Empirical therapy and prevention of infection in the immunocompromised host. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and practice of infectious diseases. 4th ed. New York: Churchill Livingstone, 1995: Benador N, Siergrist CA, Gendrel D, et al. Procalcitonin is a marker of severity of renal lesions in pyelonephritis. Pediatrics 1998; 102: Moses LE, Emerson JD, Hosseini H. Analyzing data from ordered categories. In: Bailar JC III, Mosteller F, eds. Medical uses of statistics. 2d ed. Boston: Massachusetts Medical Society, 1992: Dandona P, Nix D, Wilson MF, et al. Procalcitonin increase after endotoxin injection in normal subjects. J Clin Endocrinol Metab 1994; 79: Oberhoffer M, Stonans I, Russwurm S, et al. Procalcitonin expression in human peripheral mononuclear cells and its modulation by lipopolysaccharides and sepsis-related cytokines in vitro. J Lab Clin Med 1999; 134: CID 2001:32 (15 June) Giamarellos-Bourboulis et al.

8 13. Beaune G, Bienvenu F, Pondarré C, Monneret G, Bienvenu J, Souillet G. Serum procalcitonin rise is only slight in two cases of disseminated aspergillosis. Infection 1998; 26: Chiesa C, Panero A, Rossi N, et al. Reliability of procalcitonin concentrations for the diagnosis of sepsis in critically ill neonates. Clin Infect Dis 1998; 26: Engel A, Steinbach G, Kern P, Kern WV. Diagnostic value of procalcitonin serum levels in neutropenic patients with fever: comparison with interleukin-8. Scand J Infect Dis 1999; 31: Ruokonen E, Nousiainen T, Pulkki K, Takala J. Procalcitonin concentrations in patients with neutropenic fever. Eur J Clin Microbiol Infect Dis 1999; 18: Procalcitonin in Diagnosis of Febrile Neutropenia CID 2001:32 (15 June) 1725