Abstract
The frequent lack of microbiological documentation of infection by blood cultures (BC) has a major impact on clinical management of febrile neutropenic patients, especially in cases of unexplained persistent fever. We assessed the diagnostic utility of the LightCycler SeptiFast test (SF), a multiplex blood PCR, in febrile neutropenia. Blood for BC and SF was drawn at the onset of fever and every 3 days of persistent fever. SF results were compared with those of BC, clinical documentation of infection, and standard clinical, radiological, and microbiological criteria for invasive fungal infections (IFI). A total of 141 febrile neutropenic episodes in 86 hematological patients were studied: 44 (31%) microbiologically and 49 (35%) clinically documented infections and 48 (34%) unexplained fevers. At the onset of fever, BC detected 44 microorganisms in 35/141 (25%) episodes. Together, BC and SF identified 78 microorganisms in 61/141 (43%) episodes (P = 0.002 versus BC or SF alone): 12 were detected by BC and SF, 32 by BC only, and 34 by SF only. In 19/52 (37%) episodes of persistent fever, SF detected 28 new microorganisms (7 Gram-positive bacterial species, 15 Gram-negative bacterial species, and 6 fungal species [89% with a clinically documented site of infection]) whereas BC detected only 4 pathogens (8%) (P = 0.001). While BC did not detect fungi, SF identified 5 Candida spp. and 1 Aspergillus sp. in 5/7 probable or possible cases of IFI. Using SeptiFast PCR combined with blood cultures improves microbiological documentation in febrile neutropenia, especially when fever persists and invasive fungal infection is suspected. Technical adjustments may enhance the efficiency of this new molecular tool in this specific setting.
Febrile neutropenia is a frequent life-threatening complication in patients with hematological malignancies. Blood cultures (BC) identify a pathogen in only 20 to 30% of febrile episodes, and the rate of microbiological documentation drops to less than 10 to 15% for those receiving antibiotics at time of sampling (6, 39). The majority of episodes are thus uniquely managed based on the presence of fever and/or of a clinical site of infection. In addition, the low sensitivity of cultures for the detection of fungi is a major concern in these patients at high risk of invasive mycoses (13, 33). The diagnosis of invasive fungal infections (IFI) based on clinical, radiological, and microbiological criteria according to the European Organization for Research and Treatment of Cancer-Mycoses Study Group (EORTC-MSG) classification is often only presumptive and delayed (11). The persistence of fever despite broad-spectrum antibacterial therapy is observed in one-third of cases, and the lack of identification of the causal pathogen results in empirical modification of antibacterial therapy and adjunction of antifungal therapy (10, 18, 29). Additional tools are thus needed for the diagnosis of infection.
Molecular methods are able to rapidly detect microorganisms without incubation and despite ongoing antimicrobial therapy (26, 28, 42). Homemade multiplex or broad-range PCR assays for the detection of bloodstream pathogens in cases of sepsis or febrile neutropenia have provided variable sensitivity and specificity compared with blood cultures (5, 8, 9, 28, 31, 42, 44). Their use is limited by the lack of standardized technical procedures and commercially available systems. The LightCycler SeptiFast test (SF) (Roche Diagnostics GmbH, Mannheim, Germany) is a real-time multiplex PCR amplification assay designed to detect a broad spectrum of bacteria and fungi in human blood from nonneutropenic patients with sepsis (21, 30). The internal transcript spacer (ITS) region of the bacterial and fungal genome is the target selected for identification of 25 bloodstream pathogens. The aim of this study was to assess the appropriateness and utility of SF for the microbiological diagnosis of infection in febrile neutropenic cancer patients.
MATERIALS AND METHODS
Patients.
This prospective, observational study was conducted in the isolation ward of the Infectious Diseases Service at the University Hospital of Lausanne (Switzerland) between September 2006 and November 2007. Consecutive adult hematological patients undergoing induction or consolidation chemotherapy for acute leukemia or autologous hematopoietic stem cell transplantation were enrolled after written informed consent. The study protocol was approved by the Institutional Ethical Committee.
Clinical management.
Patients were hospitalized in positive-pressure high-efficiency particulate air-filtered isolation rooms. During the neutropenic period, no antibacterial prophylaxis was used and patients with mucositis and oral and/or gastro-intestinal tract Candida colonization received fluconazole prophylaxis. Diagnostic workup of fever and empirical antibacterial therapy were conducted according to guidelines of the Infectious Diseases Society of America (IDSA) (18). Modification of antibacterial therapy and addition of antifungal therapy were based on the clinical course (persistent fever for >72 h, clinical deterioration, and/or a new or progressing focus of infection) and microbiological reassessment (18). Whereas clinical management was routinely based on BC results, SF results were not available in real time for therapeutic decisions.
Definitions.
Standard definitions of neutropenia (neutrophil count < 500/mm3) and fever (measured once at ≥38.3°C or twice at ≥38.0°C for a 12-h period) were used (18). A new febrile episode was defined as a new onset of fever after a 72-h period of apyrexia. The etiology of febrile episodes was classified according to definitions of the International Immunocompromised Host Society (ICHS): microbiologically documented infections with bacteremia (MDI-B) or without bacteremia (MDI-NB), clinically documented infections (CDI), or fever of unexplained origin (FUO) (1). Standard definitions (CDC and WHO) were used for CDI classifications (15). Mucositis with a WHO score of >2 and diarrhea with a frequency of >8 episodes/day were considered to represent CDI (3). Invasive fungal infections (IFI) were classified according to the definitions of the EORTC-MSG (11).
Blood sampling.
Sets of blood samples for cultures (BC) and for testing with the LightCycler SeptiFast test (Roche Diagnostics GmbH, Mannheim, Germany) (SF) were drawn at the onset of fever and every 3 days in cases of persistent fever. The same procedure was applied for each new febrile neutropenic episode. Blood (10 ml) was collected in each culture bottle (Bactec Plus aerobic/F and Lytic anaerobic/F; Becton Dickinson, Sparks, MD). Each set of samplings consisted of 4 pairs (i.e., 4 aerobic and 4 anaerobic bottles [80 ml]) of blood culture samples drawn simultaneously from the central venous catheter (2 pairs of samples consisting of 2 aerobic and 2 anaerobic bottles [40 ml]) and by peripheral venipuncture (2 pairs of samples consisting of 2 aerobic and 2 anaerobic bottles [40 ml]) according to the IDSA recommendations (18). For SF assays, 3-ml samples of blood were collected in EDTA tubes (Monovette K-EDTA; Sarstedt, Numbrecht, Germany). Each set of samplings consisted of 4 tubes (i.e., 12 ml) drawn immediately after blood cultures from the central venous catheter (2 tubes [6 ml]) and by peripheral venipuncture (2 tubes [6 ml]). SF samples were sent within 1 h to the laboratory and stored at 2 to 8°C.
Microbiological analyses.
A Bactec 9240 automated blood culture system (Becton Dickinson, Sparks, MD) was used. The vials were incubated at 35°C for 5 days.
For SF assays, DNA was extracted from the EDTA whole-blood tubes within 48 to 72 h after sampling. Specimen preparation and DNA amplification and detection were performed in a dedicated laboratory according to the manufacturer's recommendations (21, 30). Each specimen included an inhibition control, and each run included positive, negative, and extraction controls. The SF analytical spectrum includes the most common Gram-positive and Gram-negative bacteria, as well as pathogenic fungi (Table 1). For coagulase-negative staphylococci and streptococci, a semiquantitative analytical cutoff value has been set by the manufacturer in order to avoid false-positive results due to contamination by colonizing skin flora (21, 30).
TABLE 1.
Analytical spectrum of the LightCycler SeptiFast test
| Gram-positive bacterial species | Gram-negative bacterial species | Fungal species |
|---|---|---|
| Staphylococcus aureus | Escherichia coli | Candida albicans |
| Staphylococcus epidermidisa | Klebsiella pneumoniae | Candida tropicalis |
| Staphylococcus haemolyticusa | Klebsiella oxytoca | Candida parapsilosis |
| Streptococcus pneumoniae | Serratia marcescents | Candida krusei |
| Streptococcus pyogenesa | Enterobacter cloacae, Enterobacter aerogenes | Candida glabrata |
| Streptococcus agalactiaea | Proteus mirabilis | Aspergillus fumigatus |
| Streptococcus mitisa | Pseudomonas aeruginosa | |
| Enterococcus faecium | Acinetobacter baumannii | |
| Enterococcus faecalis | Stenotrophomonas maltophilia |
For coagulase-negative staphylococci and streptococci, a semiquantitative analytical cutoff value has been set by the manufacturer for distinguishing between true pathogens and contaminants from the skin flora.
Data and statistical analyses.
Positive results for SF and BC were compared for neutropenic patients at the onset of fever (day 0) and during persistent fever lasting for 3 days or more (days 3, 6, 9, etc.). Discordant results were interpreted according to clinical assessment and radiological plus microbiological findings. Microorganisms recovered by BC (including coagulase-negative staphylococci) in at least one of the four pairs (at least one of the four aerobic and four anaerobic bottles) of the sets of samples drawn at the same time were classified as true pathogens according to the recommendations for management of febrile neutropenia (2, 18, 34).
In BC-negative episodes, microorganisms identified by SF were interpreted as true pathogens in the presence of at least one of the following criteria: (i) the same microorganism was recovered at the same time by cultures from a clinically relevant site other than blood (e.g., central venous catheter, urine, sputum, or bronchoalveolar lavage [BAL] fluid or from another normally sterile site) or (ii) a site of infection was documented clinically and/or radiologically in concomitance with the positive SF finding and the microorganism identified by SF was consistent with a potential pathogen according to the site of infection (28). In the absence of the criteria listed above, the significance of a positive SF result remained indeterminate. Positive SF findings for fungi were interpreted according to the EORTC-MSG diagnostic classification of fungal infections (11).
Fisher's exact test and the nonparametric Mann-Whitney rank-sum test were used for the analysis of proportions and continuous variables, respectively. A two-sided P value < 0.05 was considered statistically significant.
RESULTS
141 febrile neutropenic episodes occurred in 86 consecutive patients, with a median of 1 per patient (range, 1 to 6). A total of 237 sets of blood samples for cultures (BC) and SeptiFast PCR (SF) were obtained, with a median of 2 sets per patient (range, 1 to 8); of those sets of samples, 144 (61%) were drawn while broad-spectrum antibiotic therapy was ongoing. The characteristics of patients and febrile episodes are shown in Table 2.
TABLE 2.
Characteristics of patients and febrile episodes
| Febrile neutropenic patient characteristic (n = 86) | Valuea |
|---|---|
| Male/female | 53 (62)/33 (38) |
| Median age in yr (range) | 54 (17-71) |
| Hematological malignancies | |
| Acute myeloid leukemia/acute | |
| lymphoblastic leukemia | 37 (43)/9 (10) |
| Lymphoma | 12 (14) |
| Multiple myeloma | 22 (26) |
| Other | 6 (7) |
| Chemotherapy | |
| Induction/consolidation for acute leukemia | 45 (52) |
| Autologous HSCTb | 37 (43) |
| Other | 4 (5) |
| Median no. of days of neutropenia (range) | 11 (1-140) |
| Febrile episodes (n = 141) | |
| First febrile episodes | 85 (60) |
| Recurrent febrile episodes | 56 (40) |
| Ongoing antimicrobial therapy at | |
| onset of fever | 51 (36) |
| Microbiologically documented infections | |
| (MDI) | 44 (31) |
| Bacteremic | 35 (25) |
| Single Gram-positive bacterial species | 13 (37) |
| Single Gram-negative bacterial species | 15 (43) |
| Polymicrobial | 7 (20) |
| Nonbacteremicc | 9 (6) |
| Clinically documented infection (CDI) | 49 (35) |
| Upper/lower gastrointestinal tract | 36 (73) |
| Upper/lower airways | 6 (12) |
| Catheter/skin/soft tissues | 7 (14) |
| Fever of unexplained origin (FUO) | 48 (34) |
| Episodes with persistent fever for ≥3 days | 52 (36) |
| Invasive fungal infections | |
| (EORTC-MSG criteria) | 7 (5) |
| Probable | 2 |
| Possible | 5 |
Values represent number (percent) except where otherwise indicated.
HSCT, hematopoietic stem cell transplantation.
Four cases of enterocolitis (Clostridium difficile infection), one of pneumonia (Pneumocystis jiroveci), and four of urinary tract infections (two Escherichia coli, one Enterobacter cloacae, and one Klebsiella pneumoniae). Clostridium difficile and Pneumocystis jiroveci are not in the SF analytical spectrum. In one of the four urinary tract infections, SF detected the same microorganism (K. pneumoniae/K. oxytoca) in blood.
Febrile episodes were classified according to the ICHS definitions as follows: 35 (25%) were MDI-B, 9 (6%) were MDI-NB, 49 (35%) were CDI, and 48 (34%) were FUO (Table 2).
Diagnostic performance of blood cultures and SeptiFast blood PCR at the onset of fever (day 0).
At the onset of fever, 44 microorganisms (21 Gram-negative and 23 Gram-positive bacteria; no fungi) were detected by BC in 35/141 (25%) febrile episodes. SF identified 46 species of microorganisms (29 Gram-negative and 13 Gram-positive bacteria and 4 fungi) in 35/141 (25%) episodes. Together, BC and SF allowed the identification of 78 species of microorganisms in 61/141 (43%) episodes (P = 0.002 compared with BC or SF alone): 12 were detected by both BC and SF, 32 by BC only, and 34 by SF only (Table 3).
TABLE 3.
Microorganisms detected at the onset of fever (day 0) by blood culture and/or the LightCycler SeptiFast test (SF)a
| Species | No. of species detected by: |
||
|---|---|---|---|
| Blood culture only | Blood culture and SF | SF only | |
| Gram-negative bacteria | 11 | 10 | 19 |
| Escherichia coli | 6 | 5 | 2 |
| Klebsiella pneumoniae/K. oxytoca | 1 | 2 | 4 |
| Enterobacter cloacae/E. aerogenes | 1 | ||
| Pseudomonas aeruginosa | 1 | 11 | |
| Acinetobacter baumannii | 1 | ||
| Stenotrophomonas maltophilia | 1 | 2 | |
| Capnocytophaga gingivalisb | 1 | ||
| Sphingomonas paucimobilisb | 1 | ||
| Leptotrichia spp.b | 1 | ||
| Gram-positive bacteria | 21 | 2 | 11 |
| Staphylococcus aureus | 7 | ||
| Staphylococcus epidermidis/S. haemolyticus | 3 | 1 | 3 |
| Streptococcus mitis/S. agalactiae | 7 | ||
| Streptococcus salivariusb | 1 | ||
| Enterococcus faecium | 1 | 1 | |
| Enterococcus faecalis | 1 | ||
| Enterococcus spp. (other)b,c | 3 | ||
| Gemella spp.b | 2 | ||
| Bacillus cereusb | 2 | ||
| Bacillus spp. (other)b | 1 | ||
| Corynebacterium spp.b | 1 | ||
| Fungi | 0 | 0 | 4 |
| Aspergillus fumigatus | 1 | ||
| Candida albicans | 1 | ||
| Candida tropicalis | 2 | ||
| Total | 32 | 12 | 34 |
Microorganisms included in the SF analytical spectrum that cannot be distinguished at the species level have been coupled (e.g., K. pneumoniae/K. oxytoca, E. cloacae/E. aerogenes, S. epidermidis/S. haemolyticus, S. mitis/S. agalactiae).
Pathogens not included in the SF analytical spectrum.
Enterococcus gallinarum, Enterococcus durans, and other Enterococcus spp.
Among the 32 species of microorganisms detected by BC only, 13 (41% [10 Gram-positive and 3 Gram-negative bacteria]) were not included in the SF analytical spectrum. Among the remaining 19 isolates, 7 were Streptococcus spp., 3 were coagulase-negative staphylococci (2 of them were recovered in only 1 of 4 pairs of bottles), and 6 were Escherichia coli. A clinically or radiologically documented focus of infection was present in 18 (56%) cases (13 cases of gastrointestinal mucositis, 2 catheter-related infections, 2 pneumonias, 1 urinary tract infection), while the 14 (44%) remaining pathogens were associated with a primary bacteremia without a documented source. These 32 species of microorganisms were all considered true pathogens according to the definitions described in Materials and Methods.
Conversely, SF detected 34 species of microorganisms (11 Gram-positive and 19 Gram-negative bacteria and 4 fungi) which were not recovered by BC in 4/35 (11%) cases of MDI-B, 2/9 (22%) cases of MDI-NB, 12/49 (24%) cases of CDI, and 8/48 (17%) cases of FUO. Of the 34 species of microorganisms detected, 22 (65%) were interpreted as true pathogens, i.e., the presence of the same microorganism recovered by culture at a site other than blood (n = 1 [urine]) or of microorganisms consistent with a clinically or radiologically documented site of infection (n = 22 [2 cases of oral and 10 of gastrointestinal mucositis and 3 cases of respiratory tract, 2 of urinary tract, and 5 of catheter-related or skin or soft tissue infections). The pathogenic role of the 12/34 (35%) remaining species of microorganisms was undetermined in the absence of a consistent site of infection. Pseudomonas aeruginosa and Staphylococcus aureus accounted for 4/12 and 3/12 of those remaining microorganisms.
The number of positive SF tubes per sampling (i.e., 4 tubes) was significantly higher when the same microorganism was also identified by BC (median, 3 [range, 1 to 4]) compared with BC-negative episodes (median, 1 [range, 1 to 4; P < 0.0001]) and in cases of CDI compared with FUO (a median of 1 and a range of 1 to 4 versus a median of 1 and a range of 1 to 2; P = 0.03).
Episodes with persistent fever for ≥3 days.
Persistent fever for ≥3 days despite ongoing antimicrobial therapy was observed in 52/146 (36%) febrile episodes (19 MDI, 23 CDI, and 10 FUO). The median duration of fever in these episodes was 7 days (range, 3 to 25).
Additional species of microorganisms were detected by BC on or beyond day 3 of persistent fever in 4 (8%) episodes. In contrast, the yield of SF was significantly higher, with 28 new bacterial or fungal species identified in 19 (37%) episodes of persistent febrile neutropenia (P = 0.001). The four species detected by BC (one Stenotrophomonas maltophilia, two S. epidermidis, and one S. haemolyticus), two of which were also detected by SF, were resistant to ongoing antibacterial therapy. The 28 additional species of microorganisms detected by SF and not by BC were 15 Gram-negative bacteria, 7 Gram-positive bacteria, and 6 fungi. Twenty-five (89%) of them were identified in the presence of a consistent clinical site of infection: 2 oral and 19 gastrointestinal cases of mucositis, 3 respiratory tract infections, and 1 catheter-related or soft tissue infection (Table 4).
TABLE 4.
Microorganisms identified by the LightCycler SeptiFast test (SF) only in samples from patients with blood culture-negative neutropenic episodes and with persistent fever on day 3 or beyond 3 days after onset of fever according to the clinically and/or radiologically documented site of infection
| Site of infection | Species identified on indicated day of fever (no. of episodes) |
|
|---|---|---|
| 3 | > 3 | |
| Upper gastrointestinal tract | P. aeruginosa, S. aureus | |
| Lower gastrointestinal tract | P. aeruginosa (4), S. aureus, E. faecium, A. fumigatus, C. parapsilosis | E. coli (3), Klebsiella spp.,aEnterobacter spp.,a, P. aeruginosa, E. faecium (2), C. albicans (2), C. tropicalis |
| Lung | P. aeruginosa, S. pneumoniae | C. albicans |
| Catheter/skin | S. aureus | |
| No site | P. aeruginosa, A. baumannii | P. aeruginosa |
Klebsiella spp. and Enterobacter spp. cannot be distinguished at the species level by SF.
Invasive fungal infections (IFI).
IFI was diagnosed according to the EORTC-MSG criteria (11) in 7 (5%) febrile episodes: 2 probable and 1 possible cases of aspergillosis, 3 possible cases of candidiasis (2 of which were associated with seroconversion of antimannan antibodies), and 1 possible IFI. While no fungemia was detected by BC, SF detected fungi in 5/7 IFI cases (4 candidiasis and 1 aspergillosis) (Table 5). Fungi were detected by SF at a median of 9 days (range, 6 to 15) after the onset of fever, while IFI was diagnosed by EORTC-MSG criteria after a median of 23 days (range, 8 to 106) and empirical antifungal therapy had been started after a median of 7 days (range, 6 to 8). SF diagnosis preceded EORTC-MSG diagnosis in 3 cases. SF was negative in two cases of probable pulmonary aspergillosis, with positive bronchoalveolar lavage (BAL) galactomannan and negative blood galactomannan test results. In addition, SF was positive for fungi in 4 febrile episodes that were not classified as IFI according to EORTC-MSG criteria (Table 5). Enterocolitis (CDI) and gastrointestinal tract colonization with Candida spp. were documented in 2 cases (positive SF results for C. albicans and C. tropicalis). In the remaining 2 cases, no clinical symptoms or signs of infection were found.
TABLE 5.
Characteristics of febrile episodes with invasive fungal infections (IFI) and/or positive LightCycler SeptiFast test (SF) results for fungia
| IFI (EORTC-MSG criteria) | SF result | Persistent fever (>3 days) | Gastrointestinal tract mucositis and Candida colonization | Serological marker (GM, Mn, anti-Mn)b | CT finding consistent with IFI | Ongoing antifungal therapyc |
|---|---|---|---|---|---|---|
| Probable aspergillosis | - | + | − | Pos GMn in BALd | + (lungs) | + |
| Probable aspergillosis | - | + | − | Pos GMn in BALd | + (lungs) | + |
| Possible aspergillosis | A. fumigatus | − | − | Neg | + (lungs) | + |
| Possible candidiasis | C. albicans/C. parapsilosis | + | + | Pos anti-Mnd | + (liver, spleen) | − |
| Possible candidiasis | C. albicans | + | + | Pos anti-Mnd | + (liver, lungs) | + (empirical) |
| Possible candidiasis | C. tropicalis | + | + | Neg | + (spleen) | + (empirical) |
| Possible IFI | C. albicans | + | − | Neg | + (lungs) | + (empirical) |
| None | A. fumigatus | + | − | Neg | −e | − |
| None | C. albicans | − | + | Neg | − | − |
| None | C. tropicalis | + | + | Neg | − | + (prophylactic) |
| None | C. tropicalis | − | − | Neg | −e | − |
+, condition present; −, condition absent.
GMn, galactomannan; Mn, mannan; anti-Mn, anti-mannan antibodies; BAL, bronchoalveolar lavage fluid; Pos, positive; Neg, negative.
Values represent those seen at the time of a positive SF result or of an IFI diagnosis (EORTC-MSG criteria) when the SF result was negative.
Positive galactomannan, index ≥ 0.5; positive anti-mannan, ≥10 arbitrary units/ml.
Computed tomography (CT) not done.
DISCUSSION
The performance of the LightCycler SeptiFast test (SF) in combination with blood culture (BC) was assessed with a large sample of febrile neutropenic patients. When used together, SF and BC significantly increased the rate of microbiological documentation at the onset of fever compared with the use of BC or SF alone. While the yields of BC alone and SF alone were similar at the initial presentation of fever, SF provided a high number of positive results in cases of BC-negative persistent neutropenic fever by identifying new bacterial and fungal microorganisms consistent with a clinically documented site of infection.
Studies performed with nonneutropenic patients with sepsis showed that SF may be used in addition to BC to improve the timing and performance of the etiological documentation of infection (4, 7, 12, 14, 21-23, 27, 35, 36, 38, 41, 43). Few data are available on the utility of SF in febrile neutropenia, although recent studies have suggested that it may be a helpful supplementary tool for the detection of bacteria or fungi (24, 37, 40). BC typically fail to detect pathogens when antimicrobial therapy is ongoing. Our study identified BC-negative episodes—in particular, those occurring in patients with persistent neutropenic fever—as the clinical subset in which this molecular diagnostic tool may best complement cultures for the etiological diagnosis of bacterial and fungal infections. These data suggest that SF is an efficient tool for the early diagnosis of IFI when combined with clinical, microbiological, and radiological criteria. As the time window of SF positivity overlapped with the start of empirical antifungal therapy, SF may have an impact on therapeutic decisions. Moreover, the detection of circulating microbial DNA after prolonged appropriate anti-infective therapy in a patient with persistent signs of infection suggests a sustained release from the infectious focus. This finding and those reported from recent studies on molecular diagnosis performed in critically ill patients may provide a new insight into the natural history of infection in these high-risk settings (8, 22, 42, 43). Although SF does not provide information on the susceptibility of microorganisms, the species identification may be useful for the reassessment of the appropriateness of ongoing antimicrobial therapy when infection is not responding to treatment.
The occurrence of about one-third false-negative SF results has been reported in studies of nonneutropenic patients (22, 43). While SF failed to detect frequent bloodstream pathogens such as E. coli, coagulase-negative staphylococci, and streptococci, about 40% of the false-negative SF results in our analysis occurred with microorganisms not included in the SF analytical spectrum. The smaller blood volumes used for SF compared with BC may have contributed to the failure to detect some pathogens. The low sensitivity of SF for coagulase-negative staphylococci and streptococci is possibly associated with the semiquantitative analytical cutoff value specifically set in order to distinguish contamination from infection in nonneutropenic patients with sepsis (21). Finally, the presence of high bacterial loads might have resulted in a paradoxical inhibition of the molecular detection.
The absence of a reliable diagnostic gold standard is a common limitation for the assessment of new molecular techniques: in particular, negative BC results due to ongoing antimicrobial therapy or small circulating microbial loads and positive BC results due to contaminations represent a major limitation for the interpretation of positive or negative SF results (28). In this context, the calculation of sensitivity, specificity, positive and negative predictive values, and likelihood ratios as well as of the efficiency of the SF may be inappropriate. Experts have recommended the interpretation of positive PCR results for BC-negative febrile episodes according to the clinical context (26, 28). In the present analysis, the majority of the microorganisms detected by SF could be considered to represent the etiological cause in the presence of a clinically or radiologically documented site of infection. These results may be particularly significant for patients with persistent or progressing signs of infection, who may benefit from a modification of empirical antimicrobial therapy (4, 26, 35). Our data also suggest that the number of positive SF tubes may contribute in discriminating true-positive from false-positive results. Although the detection of P. aeruginosa and S. aureus in blood cultures usually reflects true infection, the unusually large proportion of these bacteria in a single SF tube in the absence of a documented focus may suggest a potential contamination. Despite the use of MagNA Pure water, contamination of PCR assays by waterborne bacteria such as Pseudomonas spp. is frequently reported (16, 19, 20). Although all the assay-specific internal control results were negative in our cases, low-level DNA contamination cannot be ruled out. Blood sampling from indwelling vascular catheters or by venipuncture may represent another source of contamination by cutaneous flora. On the other hand, the transient presence of viable bacteria or degraded bacterial DNA may be detected in bloodstream samples due to the breakdown of the protective barriers of skin and gastrointestinal tract mucosal membranes and/or due to the effect of ongoing efficacious antimicrobial therapy.
Although SF looks promising for the early diagnosis of IFI, the small number of probable or possible IFI and the absence of microbiologically or histopathologically proven cases are limitations of this data set that do not allow the estimation of the test's diagnostic performance. The failure of SF to detect two cases of probable pulmonary aspergillosis with negative galactomannan antigenemia might suggest that its utility for the diagnosis of localized infections is limited. These results are consistent with the limited sensitivity and specificity of blood PCR methods for the diagnosis of aspergillosis reported in previous studies (25). SF detected Candida DNA in four cases of probable or possible invasive candidiasis, while BC results remained negative. Minor differences among BC systems with respect to detection of Candida spp. do not explain the discrepancy discussed above, which is consistent with the identification by PCR of the circulation of unviable fungal components or of very low viable fungal loads (17, 32). However, not having used BC bottles containing a specific fungal medium may have potentially limited the detection of yeasts by cultures.
The SF has been designed for the detection of pathogens in septic nonneutropenic patients, and adjustments of the microbial spectrum and the analytical technique may improve its efficiency in febrile neutropenic patients (21). In particular, reassessment of the semiquantitative analytical cutoff threshold for coagulase-negative staphylococci set by the manufacturer in order to distinguish true pathogens from contaminants for intensive care unit (ICU) or emergency department patients might be appropriate for febrile neutropenic cancer patients, for whom a central line is very frequently used; the interpretations and implications for management of the recovery of coagulase-negative staphylococci from blood may substantially differ from those in other clinical settings, and the absence of neutrophils in the analyzed blood may influence the detection of microorganisms. In addition, the technique in its current version is work-intensive: automating the test protocols would facilitate its cost-effective use on a routine basis.
In conclusion, the LightCycler SeptiFast test gives new insights into the natural history of infection during neutropenia. These data suggest that this multiplex PCR technique, when combined with BC, provides clinically relevant information for the diagnosis of infection in cases of BC-negative febrile neutropenia, particularly in persistent fever despite antibacterial therapy, when a nonresponding bacterial infection or an invasive fungal infection is suspected.
Acknowledgments
We thank Cyril André, René Brouillet, Monika Ochsner, Annie Savoie, and the staff of the Isolation Ward of the Infectious Diseases Service and of the Laboratory of Microbiology and Molecular Diagnostics of the Institute of Microbiology for outstanding assistance in collection and management of clinical data and blood samples.
Financial support for this work was provided in the form of an unrestricted research grant from Roche Diagnostics GmbH, Rotkreuz, Switzerland, and Penzberg, Germany.
Potential conflict of interest: Jacques Bille, Thierry Calandra, and Oscar Marchetti attended advisory board meetings and/or gave lectures at symposia sponsored by Roche Diagnostics. All other authors declare no potential conflicts of interest.
Footnotes
Published ahead of print on 18 August 2010.
REFERENCES
- 1.Anonymous (from the Immunocompromised Host Society). 1990. The design, analysis, and reporting of clinical trials on the empirical antibiotic management of the neutropenic patient. J. Infect. Dis. 161:397-401. [DOI] [PubMed] [Google Scholar]
- 2.Beekmann, S. E., D. J. Diekema, and G. V. Doern. 2005. Determining the clinical significance of coagulase-negative staphylococci isolated from blood cultures. Infect. Control Hosp. Epidemiol. 26:559-566. doi: 10.1086/502584. [DOI] [PubMed] [Google Scholar]
- 3.Blijlevens, N., M. Schwenkglenks, P. Bacon, A. D'Addio, H. Einsele, J. Maertens, D. Niederwieser, W. Rabitsch, A. Roosaar, T. Ruutu, H. Schouten, R. Stone, S. Vokurka, B. Quinn, and S. McCann. 2008. Prospective oral mucositis audit: oral mucositis in patients receiving high-dose melphalan or BEAM conditioning chemotherapy—European Blood and Marrow Transplantation Mucositis Advisory Group. J. Clin. Oncol 26:1519-1525. doi: 10.1200/JCO.2007.13.6028. [DOI] [PubMed] [Google Scholar]
- 4.Bloos, F., F. Hinder, K. Becker, S. Sachse, D. A. Mekontso, E. Straube, V. Cattoir, C. Brun-Buisson, K. Reinhart, G. Peters, and M. Bauer. 2010. A multicenter trial to compare blood culture with polymerase chain reaction in severe human sepsis. Intensive Care Med. 36:241-247. doi: 10.1007/s00134-009-1705-z. [DOI] [PubMed] [Google Scholar]
- 5.Bretagne, S., and J. M. Costa. 2005. Towards a molecular diagnosis of invasive aspergillosis and disseminated candidosis. FEMS Immunol. Med. Microbiol. 45:361-368. doi: 10.1016/j.femsim.2005.05.012. [DOI] [PubMed] [Google Scholar]
- 6.Bucaneve, G., A. Micozzi, F. Menichetti, P. Martino, M. S. Dionisi, G. Martinelli, B. Allione, D. D'Antonio, M. Buelli, A. M. Nosari, D. Cilloni, E. Zuffa, R. Cantaffa, G. Specchia, S. Amadori, F. Fabbiano, G. L. Deliliers, F. Lauria, R. Foa, and A. Del Favero. 2005. Levofloxacin to prevent bacterial infection in patients with cancer and neutropenia. N. Engl. J. Med. 353:977-987. doi: 10.1056/NEJMoa044097. [DOI] [PubMed] [Google Scholar]
- 7.Casalta, J. P., F. Gouriet, V. Roux, F. Thuny, G. Habib, and D. Raoult. 2009. Evaluation of the LightCycler(R) SeptiFast test in the rapid etiologic diagnostic of infectious endocarditis. Eur. J. Clin. Microbiol. Infect. Dis. 28:569-573. doi: 10.1007/s10096-008-0672-6. [DOI] [PubMed] [Google Scholar]
- 8.Chan, K. Y., H. S. Lam, H. M. Cheung, A. K. Chan, K. Li, T. F. Fok, and P. C. Ng. 2009. Rapid identification and differentiation of Gram-negative and Gram-positive bacterial bloodstream infections by quantitative polymerase chain reaction in preterm infants. Crit. Care Med. 37:2441-2447. doi: 10.1097/CCM.0b013e3181a554de. [DOI] [PubMed] [Google Scholar]
- 9.Chen, S. C., C. L. Halliday, and W. Meyer. 2002. A review of nucleic acid-based diagnostic tests for systemic mycoses with an emphasis on polymerase chain reaction-based assays. Med. Mycol. 40:333-357. [DOI] [PubMed] [Google Scholar]
- 10.Cometta, A., W. V. Kern, R. De Bock, M. Paesmans, M. Vandenbergh, F. Crokaert, D. Engelhard, O. Marchetti, H. Akan, A. Skoutelis, V. Korten, M. Vandercam, H. Gaya, A. Padmos, J. Klastersky, S. Zinner, M. P. Glauser, T. Calandra, and C. Viscoli. 2003. Vancomycin versus placebo for treating persistent fever in patients with neutropenic cancer receiving piperacillin-tazobactam monotherapy. Clin. Infect. Dis. 37:382-389. doi: 10.1086/376637. [DOI] [PubMed] [Google Scholar]
- 11.De Pauw, B., T. J. Walsh, J. P. Donnelly, D. A. Stevens, J. E. Edwards, T. Calandra, P. G. Pappas, J. Maertens, O. Lortholary, C. A. Kauffman, D. W. Denning, T. F. Patterson, G. Maschmeyer, J. Bille, W. E. Dismukes, R. Herbrecht, W. W. Hope, C. C. Kibbler, B. J. Kullberg, K. A. Marr, P. Munoz, F. C. Odds, J. R. Perfect, A. Restrepo, M. Ruhnke, B. H. Segal, J. D. Sobel, T. C. Sorrell, C. Viscoli, J. R. Wingard, T. Zaoutis, and J. E. Bennett. 2008. Revised definitions of invasive fungal disease from the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group. Clin. Infect. Dis. 46:1813-1821. doi: 10.1086/588660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dierkes, C., B. Ehrenstein, S. Siebig, H. J. Linde, U. Reischl, and B. Salzberger. 2009. Clinical impact of a commercially available multiplex PCR system for rapid detection of pathogens in patients with presumed sepsis. BMC Infect. Dis. 9:126. doi: 10.1186/1471-2334-9-126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Einsele, H., and J. Loeffler. 2008. Contribution of new diagnostic approaches to antifungal treatment plans in high-risk haematology patients. Clin. Microbiol. Infect. 14(Suppl. 4):37-45. doi: 10.1111/j.1469-0691.2008.01980.x. [DOI] [PubMed] [Google Scholar]
- 14.Falagas, M. E., and P. T. Tassios. 2008. Enhanced and earlier detection of bacteremia and fungemia by multiplex polymerase chain reaction: how much enhanced, how much earlier, and at what cost? Crit. Care Med. 36:1660-1661. doi: 10.1097/CCM.0b013e31817045e5. [DOI] [PubMed] [Google Scholar]
- 15.Garner, J. S., W. R. Jarvis, T. G. Emori, T. C. Horan, and J. M. Hughes. 1996. CDC definitions for nosocomial infections, p. A1-A20. In R. N. Olmsted (ed.), APIC infection control and applied epidemiology: principles and practice. Mosby, St. Louis, MO.
- 16.Grahn, N., M. Olofsson, K. Ellnebo-Svedlund, H. J. Monstein, and J. Jonasson. 2003. Identification of mixed bacterial DNA contamination in broad-range PCR amplification of 16S rDNA V1 and V3 variable regions by pyrosequencing of cloned amplicons. FEMS Microbiol. Lett. 219:87-91. [DOI] [PubMed] [Google Scholar]
- 17.Horvath, L. L., B. J. George, C. K. Murray, L. S. Harrison, and D. R. Hospenthal. 2004. Direct comparison of the BACTEC 9240 and BacT/ALERT 3D automated blood culture systems for candida growth detection. J. Clin. Microbiol. 42:115-118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hughes, W. T., D. Armstrong, G. P. Bodey, E. J. Bow, A. E. Brown, T. Calandra, R. Feld, P. A. Pizzo, K. V. Rolston, J. L. Shenep, and L. S. Young. 2002. 2002 guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin. Infect. Dis. 34:730-751. doi: 10.1086/339215. [DOI] [PubMed] [Google Scholar]
- 19.Kawai, M., E. Matsutera, H. Kanda, N. Yamaguchi, K. Tani, and M. Nasu. 2002. 16S ribosomal DNA-based analysis of bacterial diversity in purified water used in pharmaceutical manufacturing processes by PCR and denaturing gradient gel electrophoresis. Appl. Environ. Microbiol. 68:699-704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kulakov, L. A., M. B. McAlister, K. L. Ogden, M. J. Larkin, and J. F. O'Hanlon. 2002. Analysis of bacteria contaminating ultrapure water in industrial systems. Appl. Environ. Microbiol. 68:1548-1555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lehmann, L. E., K. P. Hunfeld, T. Emrich, G. Haberhausen, H. Wissing, A. Hoeft, and F. Stuber. 2008. A multiplex real-time PCR assay for rapid detection and differentiation of 25 bacterial and fungal pathogens from whole blood samples. Med. Microbiol. Immunol. 197:313-324. doi: 10.1007/s00430-007-0063-0. [DOI] [PubMed] [Google Scholar]
- 22.Louie, R. F., Z. Tang, T. E. Albertson, S. Cohen, N. K. Tran, and G. J. Kost. 2008. Multiplex polymerase chain reaction detection enhancement of bacteremia and fungemia. Crit. Care Med. 36:1487-1492. doi: 10.1097/CCM.0b013e31816f487c. [DOI] [PubMed] [Google Scholar]
- 23.Mancini, N., S. Carletti, N. Ghidoli, P. Cichero, C. M. Ossi, R. Ieri, E. Poli, R. Burioni, and M. Clementi. 2009. Molecular diagnosis of polymicrobial sepsis. J. Clin. Microbiol. 47:1274-1275. doi: 10.1128/JCM.00011-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mancini, N., D. Clerici, R. Diotti, M. Perotti, N. Ghidoli, D. De Marco, B. Pizzorno, T. Emrich, R. Burioni, F. Ciceri, and M. Clementi. 2008. Molecular diagnosis of sepsis in neutropenic patients with haematological malignancies. J. Med. Microbiol. 57:601-604. doi: 10.1099/jmm.0.47732-0. [DOI] [PubMed] [Google Scholar]
- 25.Mengoli, C., M. Cruciani, R. A. Barnes, J. Loeffler, and J. P. Donnelly. 2009. Use of PCR for diagnosis of invasive aspergillosis: systematic review and meta-analysis. Lancet Infect. Dis. 9:89-96. doi: 10.1016/S1473-3099(09)70019-2. [DOI] [PubMed] [Google Scholar]
- 26.Nakamura, A., Y. Sugimoto, K. Ohishi, Y. Sugawara, A. Fujieda, F. Monma, K. Suzuki, M. Masuya, K. Nakase, Y. Matsushima, H. Wada, N. Katayama, and T. Nobori. 2010. Diagnostic value of PCR analysis of bacteria and fungi from blood in empiric-therapy-resistant febrile neutropenia. J. Clin. Microbiol. 48:2030-2036. doi: 10.1128/JCM.01700-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Paolucci, M., M. G. Capretti, P. Dal Monte, L. Corvaglia, M. P. Landini, S. Varani, A. Pession, G. Faldella, and V. Sambri. 2009. Laboratory diagnosis of late-onset sepsis in newborns by multiplex real-time PCR. J. Med. Microbiol. 58:533-534. doi: 10.1099/jmm.0.003848-0. [DOI] [PubMed] [Google Scholar]
- 28.Peters, R. P., M. A. van Agtmael, S. A. Danner, P. H. Savelkoul, and C. M. Vandenbroucke-Grauls. 2004. New developments in the diagnosis of bloodstream infections. Lancet Infect. Dis. 4:751-760. doi: 10.1016/S1473-3099(04)01205-8. [DOI] [PubMed] [Google Scholar]
- 29.Ramphal, R. 2004. Changes in the etiology of bacteremia in febrile neutropenic patients and the susceptibilities of the currently isolated pathogens. Clin. Infect. Dis. 39(Suppl. 1):S25-S31. doi: 10.1086/383048. [DOI] [PubMed] [Google Scholar]
- 30.Roche Diagnostics. 2006. LightCycler SeptiFast test package insert. Roche Diagnostics GmbH, Mannheim, Germany.
- 31.Rowther, F. B., C. S. Rodrigues, M. S. Deshmukh, F. N. Kapadia, A. Hegde, A. P. Mehta, and V. R. Joshi. 2009. Prospective comparison of eubacterial PCR and measurement of procalcitonin levels with blood culture for diagnosing septicemia in intensive care unit patients. J. Clin. Microbiol. 47:2964-2969. doi: 10.1128/JCM.00418-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sandven, P., L. Bevanger, A. Digranes, H. H. Haukland, T. Mannsaker, and P. Gaustad. 2006. Candidemia in Norway (1991 to 2003): results from a nationwide study. J. Clin. Microbiol. 44:1977-1981. doi: 10.1128/JCM.00029-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Segal, B. H., N. G. Almyroudis, M. Battiwalla, R. Herbrecht, J. R. Perfect, T. J. Walsh, and J. R. Wingard. 2007. Prevention and early treatment of invasive fungal infection in patients with cancer and neutropenia and in stem cell transplant recipients in the era of newer broad-spectrum antifungal agents and diagnostic adjuncts. Clin. Infect. Dis. 44:402-409. doi: 10.1086/510677. [DOI] [PubMed] [Google Scholar]
- 34.Souvenir, D., D. E. Anderson, Jr., S. Palpant, H. Mroch, S. Askin, J. Anderson, J. Claridge, J. Eiland, C. Malone, M. W. Garrison, P. Watson, and D. M. Campbell. 1998. Blood cultures positive for coagulase-negative staphylococci: antisepsis, pseudobacteremia, and therapy of patients. J. Clin. Microbiol. 36:1923-1926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Struelens, M. J. 2010. Detection of microbial DNAemia: does it matter for sepsis management? Intensive Care Med. 36:193-195. doi: 10.1007/s00134-009-1710-2. [DOI] [PubMed] [Google Scholar]
- 36.Tsalik, E. L., D. Jones, B. Nicholson, L. Waring, O. Liesenfeld, L. P. Park, S. W. Glickman, L. B. Caram, R. J. Langley, J. C. van Velkinburgh, C. B. Cairns, E. P. Rivers, R. M. Otero, S. F. Kingsmore, T. Lalani, V. G. Fowler, and C. W. Woods. 2009. Multiplex PCR to diagnose bloodstream infections in patients admitted from the emergency department with sepsis. J. Clin. Microbiol. 48:26-33. doi: 10.1128/JCM.01447-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Varani, S., M. Stanzani, M. Paolucci, F. Melchionda, G. Castellani, L. Nardi, M. P. Landini, M. Baccarani, A. Pession, and V. Sambri. 2009. Diagnosis of bloodstream infections in immunocompromised patients by real-time PCR. J. Infect. 58:346-351. doi: 10.1016/j.jinf.2009.03.001. [DOI] [PubMed] [Google Scholar]
- 38.Vince, A., S. Z. Lepej, B. Barsic, D. Dusek, Z. Mitrovic, R. Serventi-Seiwerth, and B. Labar. 2008. LightCycler SeptiFast assay as a tool for the rapid diagnosis of sepsis in patients during antimicrobial therapy. J. Med. Microbiol. 57:1306-1307. doi: 10.1099/jmm.0.47797-0. [DOI] [PubMed] [Google Scholar]
- 39.Viscoli, C., A. Cometta, W. V. Kern, R. Bock, M. Paesmans, F. Crokaert, M. P. Glauser, and T. Calandra. 2006. Piperacillin-tazobactam monotherapy in high-risk febrile and neutropenic cancer patients. Clin. Microbiol. Infect. 12:212-216. doi: 10.1111/j.1469-0691.2005.01297.x. [DOI] [PubMed] [Google Scholar]
- 40.von Lilienfeld-Toal, M., L. E. Lehmann, A. D. Raadts, C. Hahn-Ast, K. S. Orlopp, G. Marklein, I. Purr, G. Cook, A. Hoeft, A. Glasmacher, and F. Stuber. 2009. Utility of a commercially available multiplex real-time PCR assay to detect bacterial and fungal pathogens in febrile neutropenia. J. Clin. Microbiol. 47:2405-2410. doi: 10.1128/JCM.00491-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Wallet, F., S. Nseir, L. Baumann, S. Herwegh, B. Sendid, M. Boulo, M. Roussel-Delvallez, A. V. Durocher, and R. J. Courcol. 2009. Preliminary clinical study using a multiplex real-time PCR test for the detection of bacterial and fungal DNA directly in blood. Clin. Microbiol. Infect. 16:774-779. doi: 10.1111/j.1469-0691.2009.02940.x. [DOI] [PubMed] [Google Scholar]
- 42.Wellinghausen, N., A. J. Kochem, C. Disque, H. Muhl, S. Gebert, J. Winter, J. Matten, and S. G. Sakka. 2009. Diagnosis of bacteremia in whole-blood samples by use of a commercial universal 16S rRNA gene-based PCR and sequence analysis. J. Clin. Microbiol. 47:2759-2765. doi: 10.1128/JCM.00567-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Westh, H., G. Lisby, F. Breysse, B. Boddinghaus, M. Chomarat, V. Gant, A. Goglio, A. Raglio, H. Schuster, F. Stuber, H. Wissing, and A. Hoeft. 2009. Multiplex real-time PCR and blood culture for identification of bloodstream pathogens in patients with suspected sepsis. Clin. Microbiol. Infect. 15:544-551. doi: 10.1111/j.1469-0691.2009.02736.x. [DOI] [PubMed] [Google Scholar]
- 44.Zhao, Y., S. Park, B. N. Kreiswirth, C. C. Ginocchio, R. Veyret, A. Laayoun, A. Troesch, and D. S. Perlin. 2009. Rapid real-time nucleic acid sequence-based amplification-molecular beacon platform to detect fungal and bacterial bloodstream infections. J. Clin. Microbiol. 47:2067-2078. doi: 10.1128/JCM.02230-08. [DOI] [PMC free article] [PubMed] [Google Scholar]