HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Seegmüller, I. Articles by Miethke, T. Search for Related Content PubMed PubMed Citation Articles by Seegmüller, I. Articles by Miethke, T. Agricola Articles by Seegmüller, I. Articles by Miethke, T.

Sensitivity of the BacT/ALERT FA-medium for detection of Pseudomonas aeruginosa in pre-incubated blood cultures and its temperature-dependence

Institute of Medical Microbiology, Immunology and Hygiene, Technical University of Munich, Trogerstrasse 9, 81675 Munich, Germany

Correspondence Thomas Miethke Thomas.Miethke{at}lrz.tu-muenchen.de

Received November 10, 2003
Accepted April 15, 2004

The BacT/ALERT FA-medium was evaluated to detect Pseudomonas aeruginosa in pre-incubated blood samples. As published previously its predecessor the BacT/ALERT FAN-medium failed to detect P. aeruginosa in delayed entry samples. It is now reported that FA-medium tolerates a longer pre-incubation period at 36 °C, i.e. 8 h, before detection of P. aeruginosa fails in experimentally inoculated blood cultures. In clinical blood samples the frequency of false-negative results concerning P. aeruginosa was reduced from 46.9 % (FAN-medium) to 9.1 % (FA-medium). If media inoculated with P. aeruginosa are pre-incubated at room temperature for 24 h, false-negative results are not observed. Various micro-organisms (Haemophilus influenzae, Streptococcus pneumoniae, Enterobacteriaceae, Staphylococcus aureus, Enterococcus faecalis, Candida glabrata) were recognized after pre-incubation at room temperature with similar sensitivity compared to pre-incubation at 36 °C. It is concluded that FA-medium detects P. aeruginosa in delayed entry samples with increased sensitivity and pre-incubation at room temperature is superior to pre-incubation at 36 °C.

	INTRODUCTION

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
Detection of blood-stream infections by the laboratory must be fast and reliable. Therefore, automatic detection systems were developed which monitor seeded blood cultures continuously for the presence of micro-organisms in order to reduce the time to detection as much as possible. We reported previously that the former BacT/ALERT FAN-medium failed to detect the non-fermenter Pseudomonas aeruginosa in 46.9 % of all cases and Candida species in 27.3 % of all cases in delayed entry samples from the clinic (Klaerner et al., 2000). In experimentally inoculated blood cultures, the former BacT/ALERT FAN-medium was also unable to detect P. aeruginosa and Acinetobacter baumannii if pre-incubation at 36 °C exceeded 4 h. In addition, another non-fermenter, Stenotrophomonas maltophilia, was not detected if pre-incubated for at least 16 h (Klaerner et al., 2000). The manufacturer has now changed the blood culture system. The BacT/ALERT FA-medium differs from the old formula in that it additionally contains a pancreatic digest of casein, a papaic digest of soybean meal and the concentration of charcoal was reduced from 8.5 to 6.5 % (w/v). Apart from that, the volume of the blood culture system was decreased from 40 to 30 ml. Furthermore, aerobic blood cultures must not be ventilated after inoculation. These considerable changes prompted us to investigate the ability of the BacT/ALERT FA-medium to detect P. aeruginosa in delayed entry samples from the clinic as well as experimentally inoculated and pre-incubated specimens. In addition, we analysed the influence of the pre-incubation temperature, i.e. room temperature versus 36 °C, on the frequency of false-negative results.

	METHODS

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
Experimental inoculation of blood cultures.

BacT/ALERT FA-medium (bioMérieux) was used throughout the experimental study. The new medium contains a pancreatic digest of casein, a papaic digest of soybean meal, tryptic soy broth (2.0 %, w/v), brain heart infusion solids (0.1 %, w/v), sodium polyanetholesulfonate (0.05 %, w/v), pyridoxine/HCl (0.001 %, w/v), menadione (0.00005 %, w/v), haemin (0.0005 %, w/v), activated charcoal (6.5 %, w/v), L-cysteine and other complex amino acids and carbohydrate substrates, in purified water. Duplicate bottles were used for each micro-organism, temperature and time of pre-incubation. A suspension (McFarland 0.5) of the organism to be tested (Acinetobacter baumannii, P. aeruginosa, Stenotrophomonas maltophilia and Candida glabrata) was prepared, 1 ml of each organism was added to the specific bottles and the contents were thoroughly mixed. One millilitre of medium per bottle was removed immediately after inoculation and plated on blood agar plates (Columbia 5 %, Becton Dickinson) to determine the number of c.f.u. per bottle. After overnight incubation at 36 °C colony counts were performed. Control bottles were loaded into the BacT/ALERT system immediately and experimental bottles were incubated for 4, 8, 12, 16 and 24 h at 36 °C or room temperature. According to the recommendation of the manufacturer the bottles were not vented. Bottles remained in the BacT/ALERT system until they signalled positive or until 5 days had elapsed. The time from loading to detection was determined for each bottle. Bottles that failed to signal were subcultured on blood agar plates after being unloaded. In each of these cases bacterial growth was detected.

To mimic bacterial numbers present during bacteraemia McFarland 0.5 suspensions of the organisms to be tested (Enterobacter cloacae, Escherichia coli, Klebsiella pneumoniae, Haemophilus influenzae, P. aeruginosa, Enterococcus faecalis, Staphylococcus aureus, Streptococcus pneumoniae, C. glabrata) were diluted (1 : 10 000, 1 : 100 000). BacT/ALERT FA bottles were inoculated with 1 ml of each dilution and 1 ml suspension of each bottle was removed and subcultured on blood agar plates to determine the number of c.f.u. Some of the inoculated bottles were loaded into the BacT/ALERT system immediately and the remainders were pre-incubated for 4, 8, 16, 24 h at 36 °C or room temperature. The following procedure was the same as described above.

Evaluation of patient specimens.

This study was conducted from September 2001 until November 2002 at the University hospital ‘Klinikum rechts der Isar', Technical University of Munich. A total of 5360 pairs of blood cultures (aerobic and anaerobic) were obtained during the study period. Aerobic BacT/ALERT FA and anaerobic BacT/ALERT FN bottles were used on all wards of the hospital. We included anaerobic bottles in the clinical study to obtain the whole spectrum of clinically relevant bacteria. The composition of the anaerobic medium is identical to the aerobic medium except that the concentration of sodium polyanetholesulfonate is lowered to 0.044 % (w/v) and charcoal content is increased to 8.5 % (w/v). Furthermore, aerobic bottles are enriched with O₂ whereas anaerobic bottles are enriched with N₂ and aerobic bottles contain 30 ml of media whereas anaerobic bottles contain 40 ml. The time needed for the blood culture bottles to arrive in the laboratory ranged from 30 min to 24 h. Blood culture bottles are incubated at 36 °C in the clinic in case they cannot be delivered to the laboratory right away. On arrival in the laboratory, aliquots were taken from each aerobic blood culture bottle and subcultured onto adequate agar plates (chocolate agar incubated at 36 °C, 5 % CO₂). Thereafter, bottles were loaded into the BacT/ALERT system. They remained in the system until they signalled positive or until 5 days had elapsed. Bottles that failed to signal were subcultured on blood agar plates after being unloaded. For all positive bottles the bacterial species and time from loading to detection were determined. In parallel the presence of antibiotics in clinical blood culture bottles was analysed by adding a drop of the blood culture media to agar holes in Mueller–Hinton plates inoculated with spores of Bacillus subtilis ATCC 6633 (Gatermann et al., 1997). After incubation overnight the plates were checked for growth inhibition of the indicator bacteria.

	RESULTS AND DISCUSSION

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
In order to explore the limitations of the new BacT/ALERT FA-medium we inoculated the medium with 1 ml of a McFarland 0.5 solution of non-fermenters. In a former study this condition had turned out to be the most difficult for its predecessor, the BacT/ALERT FAN-medium, to cope with (Klaerner et al., 2000). Table 1 shows that a clinical isolate of P. aeruginosa was not detected within 5 days by FA-medium if pre-incubation lasted 8 h or more, while the ATCC 27853 strain remained undetected if pre-incubation lasted for 12 h or more. Similarly, a clinical isolate of A. baumannii was not detected after a pre-incubation of 8 h or more, while FA-medium failed to detect Stenotrophomonas maltophilia after a pre-incubation of 24 h. In contrast a strain of C. glabrata was always detected.

Since the number of micro-organisms added to the culture medium by far exceeded the number of micro-organisms found in blood during bacteraemia (Whimbey et al., 1984), the inoculum was adjusted accordingly in the following experiments. A McFarland 0.5 solution of each clinical isolate of P. aeruginosa was diluted 1 : 10 000 and 1 : 100 000 and 1 ml of each solution was used to seed BacT/ALERT FA-medium. An immediate subculture on blood agar revealed that the bacterial concentration was less than 7 c.f.u. ml^–1 medium in all cases, which is considered to be in the order of magnitude found during bacteraemias. Because we suspected an inter-strain variability in terms of false-negative results (see Table 1) we used four different clinical isolates of P. aeruginosa. Table 2 shows that the mean time to detection of all four strains diluted 1 : 10 000 or 1 : 100 000 without pre-incubation was 15.2 ± 1.2 and 17.4 ± 1.5 h, respectively. There were no false-negative results. However, after 16 h of pre-incubation the first false-negative result occurred. One out of two blood culture bottles failed to detect strain #2 diluted to 1 : 10 000. All other strains at both dilutions were detected. When the pre-incubation time was lengthened to 24 h, the FA-medium failed to report strains #1 and #2 in both bottles and dilutions. Both bottles seeded with strain #4 diluted 1 : 10 000 were not reported while one out of two bottles seeded with strain #4 diluted 1 : 100 000 remained false-negative. Strain #3 was always detected.

To verify the clinical relevance of these experimental findings we analysed the ability of the BacT/ALERT FA-medium to detect micro-organisms in delayed entry samples from the clinic. Therefore a clinical study was conducted from September 2001 until November 2002 at the university hospital ‘Klinikum rechts der Isar', Technical University of Munich. All routine blood specimens (n = 5360) from the clinic were subcultured on chocolate agar on arrival in the laboratory and all bottles were incubated for 5 days in the automated blood culture system as described in Methods. Altogether 964 positive samples were found. Table 3 shows that Gram-positive bacteria were all detected except for one out of 310 isolates of coagulase-negative staphylococci. Furthermore, all bacteria belonging to the group Enterobacteriaceae and Bacteriodes spp. were recognized by the BacT/ALERT system. In contrast, the system was unable to detect three isolates out of 33 of P. aeruginosa, one isolate out of three of Stenotrophomonas maltophilia and one isolate out of 41 of Candida spp.

To exclude antibiotics as the reason for these false-negative results, we checked for the presence of antibiotics in the aerobic bottles (Gatermann et al., 1997). They were present in 17/33 blood cultures containing P. aeruginosa and could not be detected in 14/33 cultures. In 2/33 cultures the result could not be interpreted. The three false-negative cases were evenly distributed in all three groups, i.e. one bottle contained antibiotics, one did not, and in one case the result could not be interpreted.

In an attempt to further decrease the occurrence of false-negative results we lowered the pre-incubation temperature to room temperature. The data in Table 1 demonstrate that even with the huge inoculum of 1 ml of a 0.5 McFarland solution of P. aeruginosa and a pre-incubation time of as long as 24 h there were no more false-negative results. This was true for a clinical isolate as well as for strain ATCC 27853. To mimic the conditions of bacteraemia we reduced the inoculum by dilution (1 : 100 000). As expected, Table 4 shows that as long as the pre-incubation is performed at room temperature the number of false-negative results is reduced to zero.

Subsequently, we checked whether this improvement in sensitivity is true for other organisms as well, especially for fastidious organisms like H. influenzae and Streptococcus pneumoniae. All experiments were performed with a 1 : 100 000 dilution of a McFarland 0.5 solution. All H. influenzae bottles were detected irrespective of the pre-incubation temperature (Table 5). Nevertheless, there was a substantial increase in detection time from a mean of 4.25 h for pre-incubation at 36 °C to a mean of 20.1 h for pre-incubation at room temperature for 24 h (Table 5). With Streptococcus pneumoniae the situation was comparable to the situation described for P. aeruginosa. Pre-incubation for 24 h at 36 °C resulted in false-negative blood culture bottles whereas all bottles pre-incubated at room temperature signalled positive (Table 5).

Finally, we analysed the influence of the pre-incubation temperature on the detection of other organisms known to cause bacteraemia, i.e. Escherichia coli, K. pneumoniae, Enterobacter cloacae, Staphylococcus aureus, Enterococcus faecalis and C. glabrata. Again, 1 ml of a McFarland 0.5 solution diluted 1 : 100 000 was used for all micro-organisms for inoculation. In short, there was not a single false-negative result. Nevertheless, we noted an increase in detection time as soon as pre-incubation was performed at room temperature. With the exception of Enterobacter cloacae, the interval from loading the bottles into the BacT/ALERT system to detection of the bottles doubled or almost tripled (Table 6).

The literature focussing on the handling of delayed entry blood cultures is scarce, but nevertheless, contradictory. According to a recommendation in the 6th edition of the Manual of Clinical Microbiology, storage of blood cultures should be performed at room temperature for a maximum of 24 h (Miller & Holmes, 1995). This was modified in the latest edition which now recommends a 2 h storage at room temperature and if a longer storage is required to follow the recommendations of the manufacturer (Thomson & Miller, 2003). In contrast, the quality guidelines published by the German Society for Microbiology (Deutsche Gesellschaft für Hygiene und Mikrobiologie, DGHM) still recommend pre-incubation at 36 °C without specifying a maximum period of time (Seifert et al., 1997). The American Society for Microbiology recommends that blood culture bottles of continuous-monitoring systems should be delivered to the laboratory as fast as possible since there is a concern that delays in transportation may cause delays in the detection of microbial growth (Dunne et al., 1997). Two major manufacturers of automated blood culture systems (Becton Dickinson and bioMérieux) recommend pre-incubation of delayed entry samples at room temperature and finally our own previous results show an influence of pre-incubation time on the performance of BacT/ALERT FAN-medium (Klaerner et al., 2000). In this study we analysed these important issues using the new BacT/ALERT FA-medium.

Our results allow two conclusions. First, the BacT/ALERT FA-medium is superior to the former one, since the frequency of false-negative results in clinical specimens with delayed entry for the detection of non-fermenters was reduced from 40.5 to 7.1 % and Candida spp. from 27.3 to 2.4 % (Klaerner et al., 2000). All other bacteria were recognized by the new medium with the same efficiency as by the former BacT/ALERT FAN-medium. The reduction of false-negative results in clinical samples correlated with our observation that FA-medium, in contrast to FAN-medium, tolerated a pre-incubation period of 4 h in experimentally inoculated blood cultures without false-negative results (Klaerner et al., 2000). The reason for the remaining false-negative results with BacT/ALERT FA-medium is still not clear. One of the false-negative specimens in the clinical study contained antibiotics, but this was also observed in samples which signalled positive. Therefore it is unlikely that the mere presence of antibiotics is the sole reason for the false-negative results.

Nevertheless, there is one obvious correlation between some variables. The longer the pre-incubation period, and the higher the inoculum, the more likely a false-negative result (Tables 1 and 2). This implies that during pre-incubation one or more ingredients of the bottle are consumed. Taking into account that the pathogen most often associated with false-negative results is P. aeruginosa, this limiting factor might be oxygen. According to the manufacturer the atmosphere in the BacT/ALERT FA-bottle is enriched with oxygen and therefore the bottle does not have to be vented. Thus, a lack of oxygen seems to be an unlikely explanation for the false-negative results. Additionally, in our previous study we reported that even with BacT/ALERT FAN-bottles (bottles which had to be vented for optimal performance) repeated ventilation could not overcome the deficiencies of the medium (Klaerner et al., 2000).

Second, pre-incubation of delayed entry samples at room temperature seems preferable since the number of false-negative results was reduced to zero in experimentally inoculated cultures. The sensitivity of the automated system appears not to be impaired by pre-incubation at room temperature, because all micro-organisms tested were detected. However, it must be noted that the time to detection of clinically important bacteria known to be a frequent cause of bacteraemia and sepsis, like Escherichia coli (36 °C, 2.0 h; RT, 3.7 h; Table 6) and Staphylococcus aureus (36 °C, 2.0 h; RT, 5.7 h; Table 6), doubles or almost triples upon pre-incubation at room temperature. An increase in detection time has to be noted for Enterococcus faecalis, H. influenzae and in part for Streptococcus pneumoniae as well. Whether the increased time to detection of about 2–4 h in the case of Escherichia coli and Staphylococcus aureus is clinically relevant, is in our opinion rather doubtful. It appears to be more important to avoid false-negative results completely.

A previous study reported partly similar results using other automated detection systems. Thus, Streptococcus pneumoniae also remained undetected by the Bactec 9240 and the Difco ESP blood culture systems when pre-incubation was performed at 35 °C (Chapin & Lauderdale, 1996). In addition, a number of other species including P. aeruginosa failed to be recognized by the Difco ESP system if pre-incubated at 35 °C (Chapin & Lauderdale, 1996). Interestingly, the authors of that study recommended to pre-incubate blood cultures at 35 °C if the delay does not exceed 8 h or 24 h for the Difco ESP and Bactec 9240 System, respectively. If the delay exceeds these time thresholds then pre-incubation should be performed at room temperature. Possibly, this would also be the optimal solution for the issue of delayed entry samples for the system analysed here. In practical terms, however, it is probably difficult to introduce these rather complicated pre-incubation rules into a daily routine. Taking into account all the results mentioned above, pre-incubation at room temperature of delayed entry samples appears to be the best solution.

	REFERENCES

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES