• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jcmPermissionsJournals.ASM.orgJournalJCM ArticleJournal InfoAuthorsReviewers
J Clin Microbiol. Feb 2000; 38(2): 513–520.
PMCID: PMC86136

Rapid Identification of Bacteria from Positive Blood Cultures by Fluorescence-Based PCR–Single-Strand Conformation Polymorphism Analysis of the 16S rRNA Gene

Abstract

Bacteremia continues to result in significant morbidity and mortality, particularly in patients who are immunocompromised. Currently, patients with suspected bacteremia are empirically administered broad-spectrum antibiotics, as definitive diagnosis relies upon the use of blood cultures, which impose significant delays in and limitations to pathogen identification. To address the limitations of growth-based identification, the sequence variability of the 16S rRNA gene of bacteria was targeted for rapid identification of bacterial pathogens isolated directly from blood cultures using a fluorescence-based PCR–single-strand conformation polymorphism (SSCP) protocol. Species-specific SSCP patterns were determined for 25 of the most common bacterial species isolated from blood cultures; these isolates subsequently served as a reference collection for bacterial identification for new cases of bacteremia. A total of 272 blood-culture-positive patient specimens containing bacteria were tested. A previously determined SSCP pattern was observed for 251 (92%) specimens, with 21 (8%) specimens demonstrating SSCP patterns distinct from those in the reference collection. Time to identification from blood culture positivity ranged from 1 to 8 days with biochemical testing, whereas identification by fluorescence-based capillary electrophoresis was obtained as early as 7 h at a calculated cost of $10 (U.S. currency) per specimen when tested in batches of 10. Limitations encountered included the inability to consistently detect mixed cultures as well as some species demonstrating identical SSCP patterns. This method can be applied directly to blood cultures or whole-blood specimens, where early pathogen identification would result in a timely diagnosis with possible implications for patient management costs and the mortality and morbidity of infections.

In the United States, there are approximately 250,000 episodes of nosocomial bloodstream infections annually (2). If community-acquired infections are also considered, the number of annual bloodstream infections approximately doubles (22). Advances in chemotherapeutic regimens, more frequent bone marrow transplant procedures, and the continuing emergence of human immunodeficiency virus disease have all contributed to the rise in the number of immunocompromised patients (1, 21). Immunocompromised patients provide a greater opportunity for microbes, particularly opportunistic pathogens, to cause disease. Opportunistic infections add to the economic burden of health care as a result of treatment and the attendant morbidity (17). The crude mortality rate in patients with bacteremia ranges from 25 to 50%, with close to one-third of deaths being directly attributable to these infections (14, 16, 17, 21). Studies have shown that appropriate antimicrobial therapy is associated with a lower attributable mortality rate when administered early (10.4%) than when given once blood culture results are known (25.8%) (22), suggesting that this is one of the most important factors contributing to a favorable outcome (4, 22). Due to the delay in the availability of blood culture results, patient management still relies primarily on clinical diagnosis. Initial empiric treatment is based on the probability of the most likely organisms causing infections and is successful in treating many infections. However, there has been a dramatic increase in acquired resistance to antibiotics in most common pathogens, and empiric antibiotic selections are increasingly complicated and expensive. Furthermore, heavy use of broad-spectrum empiric antibiotics is an important contributor to increasing acquisition of resistance.

Other limitations of conventional blood culture techniques include low sensitivity (18). Several factors contribute to this problem, including inadequate blood volumes, the presence of antibiotics in patient sera, and the fastidious characteristics of certain species not satisfied by standard blood culture systems (12, 13, 18). Furthermore, the addition of sodium polyanetholesulfate (SPS), a common additive used to increase bacterial growth by inactivating complement, has been shown to inhibit the growth of some organisms, including Neisseria meningitidis, N. gonorrhoeae, Francisella tularensis, and Moraxella catarrhalis (18). Despite improvements made in overcoming the problems associated with blood cultures, the time to detection and identification is too lengthy, normally ranging from 2 to 5 days for most organisms or longer for fastidious organisms (18). Given the advantages of early diagnosis for the outcome of infection, it is important to develop rapid, sensitive, and specific methods of bacterial and fungal identification.

Molecular diagnosis is playing an increasingly important role in the rapid detection and identification of pathogenic organisms in clinical samples. The genetic variation of ribosomal genes in bacteria offers an alternative to culturing for the detection and identification of these organisms. These genes, such as the 16S rRNA gene, demonstrate conserved sequence regions ideal for primer targeting as well as regions of variability useful for species identification (24). Sequencing of the 16S rRNA gene of an organism for the purpose of identification holds great promise because of its accuracy but, at present, is not recommended for routine identification due to cost and time constraints. Single-strand conformation polymorphism (SSCP) is a relatively recent method that was developed in 1989 (15) and that is now commonly used as a tool for mutation detection for many genes amplified by PCR. Once denatured, the single-stranded DNA fragments adopt a certain conformation based on the DNA sequence and maintain this configuration throughout electrophoresis with a nondenaturing gel. This situation results in a variation in the electrophoretic mobility of PCR products that have similar sizes but different sequences, thus allowing them to be differentiated at the detection point without being fully sequenced.

In theory, since each bacterial species has a unique 16S rRNA sequence, all organisms can be differentiated from each other using PCR-SSCP. Widjojoatmodjo et al. (23) have previously shown this method to a promising option for use in molecular diagnosis. They have performed SSCP on PCR-amplified 16S ribosomal DNA fragments using an automated slab gel sequencing system and, with few exceptions, have obtained different peak patterns for 47 bacterial species spanning a broad range of gram-negative and gram-positive organisms of clinical interest. With the advent of capillary electrophoresis, this method may possibly be used in a time- and cost-effective manner for the identification of microorganisms in clinical specimens. We have developed a PCR-SSCP method based on that of Widjojoatmodjo et al. (23); this method is rapid and sensitive and has the capability to identify all species present in clinical specimens. We have applied this technique to positive blood cultures obtained within our laboratory to determine the specificity of the method in comparison with conventional methods and to evaluate the effects of mixed cultures upon preferential amplification and detection.

MATERIALS AND METHODS

Bacterial strains.

The 25 most common blood culture isolates were identified by searching the clinical microbiology database of our institution, the Health Sciences Centre, from 1992 to 1997. The Health Sciences Centre is a 900-bed, tertiary-care teaching hospital affiliated with the University of Manitoba. For determination of the SSCP patterns of these isolates, we used American Type Culture Collection (ATCC) strains as controls in most instances and previously identified clinical blood culture isolates (Table (Table1).1). All organisms were subcultured onto Trypticase soy–5% sheep blood agar and incubated at 37°C for 24 or 48 h prior to DNA extraction. Biochemical testing of all blood culture isolates was done by conventional manual and automated methods. Manual methods included biochemical testing as suggested in the Manual of Clinical Microbiology (5). Automated analysis consisted of using MicroScan (Dade, West Sacramento, Calif.) panels for gram-negative bacilli and gram-positive cocci.

TABLE 1
Determination of SSCP peak patterns of control strainsa

The patient specimens used for the molecular identification of bacteria by SSCP were all blood cultures that became positive from 8 December 1997 to 30 March 1998. Our laboratory uses the BacT/Alert blood culture system (Organon Teknika Corp., Durham, N.C.), which is a continuously monitoring, colorimetric CO2 microbial growth detection system (20). As part of the routine identification protocol, Gram staining was performed on all blood cultures that became positive. All specimens containing exclusively fungi were omitted from the study. The study was conducted in a double-blind fashion, in which the hospital clinical laboratory technologist on duty proceeded with routine conventional testing and results were made known to our study group only after all specimens had been identified by PCR-SSCP.

DNA isolation.

The isolation of DNA from culture colonies for PCR amplification was performed with a QIAamp Tissue Kit (Qiagen, Santa Clarita, Calif.) in accordance with the manufacturer's protocol. For the isolation of bacterial DNA from blood cultures, benzyl alcohol-guanidine hydrochloride organic extraction was used as previously described (6). Basically, 100 μl of blood culture fluid was vortexed with an equal volume of lysis buffer (5 M guanidine hydrochloride, 100 mM Tris [pH 8.0]). Four hundred microliters of sterile distilled water and 800 μl of benzyl alcohol were added, and the samples was mixed, end over end, for 5 min. The samples were centrifuged for 10 min at 17,900 × g, and 400 μl of the aqueous (top) layer was removed into another 1.5-μl microcentrifuge tube. DNA precipitation occurred with the addition of 40 μl of 3 M sodium acetate and 440 μl of isopropanol, followed by centrifugation at 4°C for 15 min. The supernatant was removed, and the pellet was washed with 1 ml of 70% ethanol, air dried, and resuspended in 100 μl of water. Routinely, 10 μl of the DNA lysate was used for PCR.

Multiplex PCR protocol using fluorescent primers.

The 16S rRNA universal primers used in the multiplex PCR were as follows: 5′HEX-13B, 5′-AGG CCC GGG AAC GTA TTC AC-3′ (19); 6-FAM-RW01, 5′-AAC TGG AGG AAG GTG GGG AT-3′ (8); 5′HEX-806R, 5′-GGA CTA CCA GGG TAT CTA AT-3′ (19); and 6-FAM-515, 5′-TGC CAG CAG CCG CGG TAA-3′ (19) (Life Technologies, Burlington, Ontario, Canada). A 50-μl PCR mixture contained 10 μl of DNA template, 2.5 mM MgCl2 buffer, 0.5 μM nucleotide (total concentration) (Amersham Pharmacia Biotech, Baie d'Urfé, Québec, Canada), 1 μM each of four primers, and 2.5 U of Taq DNA polymerase (Amersham Pharmacia Biotech). The PCR was performed using a Perkin-Elmer GeneAmp PCR System 9600 with 1 cycle at 94°C for 5 min and 30 cycles at 94°C, 55°C, and 72°C for 1 min each; the mixture was incubated at 72°C for 10 min for final extension and kept at 4°C until further processing.

SSCP analysis.

Sample preparation for capillary electrophoresis involved the addition of 1 μl of diluted PCR product to the capillary electrophoresis mixture (10.5 μl of deionized formamide, 0.5 μl of 3 M NaOH, and 0.5 μl of GeneScan-500 [ROX] Size Standard (PE Applied Biosystems, Foster City, Calif.). A 1:25 dilution of the PCR product in TE buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8.0]) to a concentration of approximately 20 ng/μl resulted in ideal peak intensity analyzable on the instrument relative to the internal standard peaks. The capillary sample mixture was denatured for 2 min at 95°C and rapidly cooled on ice prior to loading of the instrument. Subsequent preparation, such as setup of the ABI PRISM 310 Genetic Analyzer, was done in accordance with the manufacturer's instructions (PE Applied Biosystems). The nondenaturing polymer matrix used was 4% SSCP polymer (4% GeneScan polymer [PE Applied Biosystems])–10% glycerol–Tris-borate-EDTA. A capillary (47 cm by 50 μm [inner diameter]) was installed, and electrophoresis conditions were set on the instrument at a 10-s injection time, a 7-kV injection voltage, a 13-kV electrophoresis voltage, a 120-s syringe pump time, a constant temperature of 30°C, and a 24-min collection time. The matrix file (GS POP-4 SSCP module and filter set A with the above conditions) was created in accordance with the manufacturer's instructions to account for spectral overlap of the various fluorescent molecules under the specific conditions used for this application. Successful analysis was derived from comparisons of the sample peak (labeled HEX or FAM) retention time to that assigned to the ROX-labeled internal standard peaks using ABI PRISM 310 GeneScan Analysis Software (PE Applied Biosystems). Staphylococcus aureus ATCC 25923 and Escherichia coli ATCC 25922 positive controls and a water negative control were processed for PCR amplification and SSCP analysis along with each batch of samples tested. The positive controls were placed at the end of a run on the ABI PRISM 310 instrument to ensure that adequate instrument conditions were maintained and therefore were providing reproducible results. The negative control, in addition to ensuring lack of contamination, served to identify small peaks considered background upon SSCP analysis.

Time and cost analyses.

The time to conventional and molecular identification of bacteria present in blood culture samples was determined from the time at which the blood culture detection system (BacT/Alert) gave a positive signal. This time was given in days for conventional testing and in hours for molecular testing.

The cost of all reagents used and labor costs were determined for conventional testing and molecular testing. Three basic algorithms for workup were represented for conventional testing: anaerobes, gram-positive cocci, and gram-negative bacilli. Molecular testing cost was determined for SSCP identification and included DNA extraction, PCR, and SSCP on the ABI PRISM 310 instrument.

RESULTS

Creation of an SSCP peak pattern database using common blood culture isolates.

The SSCP pattern of the most common organisms isolated from blood cultures at our institution from 1992 to 1997 was determined. All included organisms constituted >85% of the total organisms isolated (data not shown). Other groups of common organisms, such as diphtheroids or bacilli, were not included, as they consist of many species. All ATCC control strains were initially tested three times to ensure reproducibility, and other strains then were tested at least once to ensure intraspecies reproducibility. The electropherograms in Fig. Fig.11 show the SSCP pattern of the five bacterial species most commonly isolated from blood cultures, as analyzed with ABI PRISM 310 GeneScan Analysis Software. A total of 25 SSCP patterns were determined for the initial database prior to application of this method to unknown blood culture isolates (Table (Table1),1), each one representing each of the different bacterial species tested. The SSCP pattern was defined by comparison to an internal standard of four values determined with the GeneScan program for each of the four single-stranded PCR-amplified fragments upon nondenaturing electrophoresis.

FIG. 1
SSCP peak patterns of the five most commonly isolated blood pathogens, as analyzed with the ABI 310 PRISM Genetic Analyzer. The multiplex PCR amplification was performed using two sets of fluorescent primers (6-FAM-RW01 and 5′HEX-13B; 6-FAM-515 ...

No intraspecies variability was detected. However, some species were found to have identical patterns: S. aureus, S. hominis, S. haemolyticus, and S. auricularis; Streptococcus pneumoniae and S. mitis; and Enterobacter cloacae and Klebsiella pneumoniae. Unique to the S. pneumoniae or S. mitis pattern, one of the peaks (806R) was flat and wide, spanning a value of 350 to 370. Multiplex PCR of four strains of Propionibacterium acnes did not result in the amplification of the 6-FAM-515–5′HEX-806R fragment.

Molecular and conventional identification of blood-culture-positive specimens.

All blood culture specimens determined to be positive for bacteria from the primary Gram stain from December 1997 to March 1998 (n = 272) were processed for PCR amplification. The amplicons were detected by capillary electrophoresis after fluorescent primers were used for amplification. Of 272 positive blood cultures, a previously established SSCP pattern from Table Table11 was identified for 251 of them (92%) (Table (Table2).2). In six of those cases, a new species represented by a known SSCP pattern from the original database was indicated by conventional testing. These were closely related species for which the clinical significance was negligible, for example, Micrococcus, Gemella, and Stomatococcus spp.; P. acnes; Propionibacterium spp.; and diphtheroids.

TABLE 2
Blood culture organisms corresponding to predetermined SSCP patterns (n = 251)

In one instance, the SSCP pattern determination did not correspond to the biochemical identification. One organism had an SSCP pattern corresponding to S. salivarius, whereas the biochemical identification indicated that this organism was S. mutans. The SSCP pattern of the control S. mutans strain was distinct from that of S. salivarius. We believe that the organism is a viridans group Streptococcus species belonging to either the S. mutans or the S. salivarius group, which comprises several species.

Twenty-one isolates tested resulted in the detection of a new SSCP pattern not previously determined and therefore not present in the SSCP database. These included less common bacteremic organisms, such as Bacteroides fragilis, or those belonging to a common group, such as diphtheroids or Corynebacterium spp. These are indicated in Table Table33 along with the SSCP pattern data.

TABLE 3
Identification by conventional methods of isolates demonstrating new SSCP patterns

Multiple new SSCP patterns were found for organisms identified by conventional methods as E. cloacae. Five specimens conventionally identified as E. cloacae demonstrated new SSCP patterns. Of those, three resulted in a first new pattern, one resulted in a second new pattern (Table (Table3),3), and one had two SSCP patterns corresponding to both the E. cloacae or K. pneumoniae group and the first new pattern. The latter was identified by conventional methods as a mixed specimen containing K. pneumoniae and E. cloacae.

Identification in mixed cultures.

Of the 272 blood culture specimens tested, 40 (15%) contained more than one species, as determined either by conventional identification or by the presence of multiple SSCP patterns (Table (Table4).4). Twenty-nine specimens (72.5%) showed only one detectable SSCP pattern. Of those, one specimen contained two species that had identical SSCP patterns. Multiple SSCP patterns were discernible for 11 mixed specimens (27.5%). In two of those specimens, only one species was detected by conventional identification. Generally, gram-negative organisms were detected in a mixture of gram-positive and gram-negative organisms, and Streptococcus and Enterococcus faecalis organisms were detected in a mixture with staphylococci and other gram-positive organisms.

TABLE 4
Mixed cultures

Time to identification from blood culture positivity.

The time to final identification of blood culture isolates by conventional methods was determined to be 1 to 8 days from the time when the BacT/Alert system gave a positive signal. The proportions of organisms identified biochemically were 57.5% within 2 days and 78% within 3 days, and the remaining took 4 days or longer. These times did not include the number of days of incubation prior to the positive signal: 1 to 7 days (data not shown). The time to identification with the GeneScan program for SSCP was from 7 h (same day) to the next day for 48 specimens, including DNA extraction, PCR, capillary electrophoresis, and analysis.

Cost analysis: conventional versus molecular methods.

The cost analysis was determined for all reagents and labor for conventional testing and molecular testing. The average cost for conventional identification per blood culture isolate ranged from $39 to $45 (U.S. currency). With molecular testing, cost is significantly lower when more than one test is processed at the same time. The cost of processing one sample alone would represent the maximum cost per sample, which was calculated to be $21 (U.S. currency). As the number of positive blood cultures obtained in our laboratory averaged 10 per day, we have calculated the cost of running 10 samples for SSCP identification to be $10 (U.S. currency) per sample.

DISCUSSION

We have used DNA amplification technology as a more rapid, specific, and cost-effective way to identify bacteremia in an attempt to improve upon the impediments associated with blood culture testing. The ultimate objective of the project is to use the method developed in this study for the identification of bacteremia directly from patient blood. In this study, we have tested the method using positive clinical blood cultures to determine its utility for organism identification. This was a double-blind study in which both the technologist performing routine conventional testing and our research group were unaware of each other's interpretation until all specimens had been identified by both methods. The comparison of PCR-SSCP identification and conventional blood culture testing results for the 272 blood-culture-positive specimens was performed at the end of the study.

The bacterial 16S ribosomal gene demonstrates species-specific sequence variability that results in a unique DNA fragment conformation easily demonstrated by SSCP. We used the ABI PRISM 310 Genetic Analyzer, an automated capillary electrophoresis sequencer, to determine the species for the fluorescence-tagged bacterial amplicons with ABI PRISM 310 GeneScan Analysis Software. PCR-SSCP patterns are detected and analyzed by attributing values to each peak represented by a fluorescence-tagged single-stranded PCR product on the electropherogram relative to the internal standard, to which values were previously assigned based on the retention time. The advantages of using such a system include speed, where each run takes less than 0.5 h, the elimination of slab gels, and the elimination of radioactive detection or silver staining. Furthermore, minimal labor is required, as this “walk-away” system is highly automated and can process 1 to 48 samples after one setup. The sensitivity of fluorescence-based detection, along with the rapidity and ease of analysis, renders this system ideal for molecular diagnosis. Multihead capillary systems will become available to permit the simultaneous analysis of multiple samples.

Culture broth fluid inoculated with blood is not normally considered an ideal clinical specimen for PCR amplification, as SPS, a common additive to blood cultures that is used to enhance bacterial growth by inactivating complement, is inhibitory to the amplification reaction (6). We used a DNA extraction method for blood cultures that included a benzyl alcohol step to remove SPS from the blood culture aliquots to a level where PCR amplification was successful.

We chose the 6-FAM-515–5′HEX-806R pair and the 6-FAM-RW01–5′HEX-13B pair for SSCP due to their relatively small fragment sizes. Sequence-based conformational differences are greater with SSCP if the fragment is short (~200 to 300 bp) (9). Also, we limited the number of primer pairs used for PCR-SSCP to two, as we anticipated the detection of more than one species in mixed cultures, complicating pattern interpretation. Although the 16S rRNA V2 region reveals the greatest interspecies sequence variability, Widjojoatmodjo et al. (23) have observed with certain species instances of intraspecies polymorphism which may be difficult to interpret; therefore, we did not use this region for SSCP analysis. However, the V2 region may still be very useful in instances where the other two primer pairs result in identical patterns for multiple species.

A total of 251 of 272 blood-culture-positive specimens (92%) resulted in the detection of organisms presenting an SSCP pattern previously determined with control strains. Of these, 245 were correctly identified to the species level or as coagulase-negative staphylococci (CoNS) or viridans group streptococci. The remaining six specimens were identified conventionally as a new, but closely related species. These included Enterobacter amnigenus (E. cloacae pattern); Gemella sp. and Stomatococcus sp. (Micrococcus luteus pattern); and Propionibacterium sp. and two diphtheroids (P. acnes pattern). This information indicates that closely related species can have identical SSCP patterns.

A total of 21 of 272 specimens (8%) contained organisms presenting new SSCP patterns. To confirm that the new SSCP patterns in fact belong to the conventionally identified species, a corresponding control strain would need to be tested. These new patterns would further complete the SSCP database, providing a larger number of species that could be identified using this method.

Forty of 272 specimens contained mixed cultures. In theory, all organisms present, even in mixed cultures, are subject to amplification. Although the analysis of additional peaks for two or more organisms could become difficult, it was actually quite easy to interpret, as seen for 11 of our mixed-culture specimens. However, in 72.5% of our mixed cultures, only one SSCP pattern was detected. In the cases with mixed gram-positive and gram-negative organisms, gram-negative organisms were amplified preferentially over gram-positive organisms. Preferential amplification also occurred in cultures containing exclusively gram-positive bacteria. Causes of these results may include variable concentrations of the organisms or primer-template mismatch (3). The exact number of organisms required to obtain a detectable signal is unclear. Factors such as the number of copies of the 16S rRNA gene present in different bacterial species or the efficiency of DNA extraction from different bacteria (gram positive versus gram negative) can influence the number of cells detectable. For mixed infection, usually only one pattern is detectable. When more than two patterns are detectable, the detected peaks are of equal intensity for the two species. Mixed infections are recognizable only if two sets of peaks are detected for at least one primer pair.

Of all isolates that were undetected by PCR-SSCP in 29 mixed-culture specimens demonstrating one SSCP pattern only, two-thirds were considered contaminants. Perhaps these were not successfully amplified due to their low concentration in the specimen itself. Alternatively, contamination might have occurred following subculturing of blood media. Preferential amplification in mixed-culture specimens must be studied further to determine the significance of the nonamplified organisms.

Issues of minor discordance occurred between conventional testing and PCR-SSCP. These included CoNS that were identified as a particular species by conventional methods, while the SSCP pattern showed that they could not be that CoNS species. Their biochemical reevaluation using MicroScan panels indicated CoNS species of low probability (~40 to 73% probability), signifying that the organism was CoNS but with an incorrect species identification. Previous studies have shown that MicroScan panels accurately identify species of CoNS only 50 to 90% of the time (10). Since the clinical importance of reporting the specific species of these organisms is limited, we did not pursue this matter further.

We have observed three unique SSCP patterns for isolates identified biochemically as E. cloacae. Whether these strains represent a nontypical E. cloacae SSCP pattern resulting from isoconformers of PCR fragments or in fact another Enterobacter species remains to be determined. McLaughlin et al. have performed a phylogenetic study on numerous E. cloacae-like clinical isolates and have shown significant sequence differences in the 16S rRNA genes of E. cloacae (ATCC type strain) and clinical strains identified as E. cloacae by conventional methods; these results suggest the possibility of a new species or genus not differentiated by conventional methods (I. J. McLaughlin, N. M. Ellis, D. Chapman, M. K. Hopkins, L. Weigel, and D. E. Dodge, Abstr. 99th Gen. Meet. Am. Soc. Microbiol., abstr. R19, 1999). Sequencing of the 16S rRNA genes of the E. cloacae-like isolates in this study would help to determine whether these organisms are different from E. cloacae, as biochemical identification of this genus remains inconclusive.

The cost analysis demonstrated significant cost savings using SSCP identification: $10 (U.S. currency) per specimen when processed in a batch of 10, compared to $40 to $45 (U.S. currency) per blood culture. Furthermore, its cost-effectiveness not only pertains to materials and labor but also can be seen as a result of early administration of appropriate treatment when a patient is bacteremic, thus reducing the mortality and morbidity rates, or the reduction of empiric therapy in patients who have no systemic infections. However, DNA amplification techniques do not establish antibacterial susceptibilities, which must still be determined by conventional methods necessitating organism growth. We believe that this problem will be overcome in the future, as more and more antibiotic resistance markers are discovered and identified by PCR methods.

The molecular methods established here, in contrast to conventional technology, have several advantages. First, molecular identification is impartial to phenotypic characteristics and does not discriminate between organisms based on their growth needs or even growth itself. Species identification problems are often encountered with organisms such as CoNS and gram-positive bacilli when MicroScan panels and other automated biochemical systems are used. Molecular techniques may require minimal technological expertise, in contrast to more subjective interpretations. Conventional biochemical testing may successfully identify the majority of pathogenic bacteria. Nevertheless, the final results do not have an impact on initial patient management due to the delayed turnaround time.

We have found SSCP to be a highly successful method of identification, in that once a pattern is determined for an organism, that organism will always be recognizable by this pattern. We can further improve SSCP as an identification system by adding an extra set of primers. However, the number of peaks may become too overwhelming to identify species in a mixed culture.

Most bacterial species have a unique SSCP pattern. Unfortunately, two important pathogens, S. aureus and S. pneumoniae, cannot be differentiated from others of the same genus (CoNS and S. mitis, respectively). If a third 16S rRNA primer pair cannot discriminate between these, other potential PCR-SSCP targets, including the 5S, 23S, or intergenic regions, or other genes known to show interspecies variability, such as the HSP60 (7) or recA (11) gene, may be added to the protocol.

The time to identification from specimen collection can be as little as 6.5 h. In contrast, conventional blood culture testing takes, on average, 2 to 5 days before positive bacterial isolation. Timing is crucial given the importance of early diagnosis from a clinical point of view. The isolates that took longest for final identification by conventional methods were mixed cultures, followed by CoNS, Micrococcus, and the viridans group streptococci, among others. Those that normally took only 1 to 2 days from a blood culture-positive signal were mainly E. cloacae, E. coli, E. faecalis, E. faecium, S. aureus, and S. pneumoniae.

The ultimate goal of using molecular technology for diagnosis of infection is for the early detection of pathogens. Patient outcome is greatly improved with the early administration of appropriate antibiotic therapy (22). Empiric therapy is then reduced, helping lead to the attenuation of the emergence of antibiotic-resistant organisms. In this study, identification by PCR-SSCP was obtained from the same day to the next morning. The goal of the method involves applying this method to clinical specimens as they are collected directly from patients, as opposed to waiting for a blood-culture-positive signal. Furthermore, the increased sensitivity of PCR may detect more cases of bacteremia, particularly with patients on antimicrobial therapy (12). When the issues associated with this method, such as sensitivity and mixed infections, have been resolved, there are two ways in which the technique may be fitted into the work flow: (i) PCR can be performed on the blood at the time of detection, or (ii) direct amplification can be performed on blood specimens when they are drawn. The sensitivity of the latter needs to be studied in a clinical setting.

In summary, we have developed a molecular method for the comprehensive identification of bacterial infections. This technique is sensitive, specific, rapid, and cost-effective. Molecular diagnostic research continues to be a priority in the health care system as researchers strive to develop methods that will have a positive impact on patient care. The use of this information, together with a rapid detection system performing fluorescence-based PCR-SSCP, has the potential to overcome many of the present limitations of conventional technology, leading to the eventual use of this method in clinical laboratories.

REFERENCES

1. Bow E J. Approach to infection in patients receiving cytotoxic chemotherapy for malignancy. In: Hall J B, Schmidt G A, Wood L D H, editors. Principles of critical care. 2nd ed. New York, N.Y: McGraw-Hill Book Co.; 1998. pp. 747–771.
2. Centers for Disease Control and Prevention. Advance report of final mortality statistics, 1989. Mon Vital Statistics Rep. 1992;40:1–13.
3. Edwards M, Gibbs R A. Multiplex PCR. In: Dieffenbach C W, Dveksler G S, editors. PCR primer: a laboratory manual. Cold Spring Harbor, N.Y: Cold Spring Harbor Laboratory Press; 1995. pp. 157–171.
4. Elting L S, Rubenstein E B, Rolston K V I, Bodey G P. Outcomes of bacteremia in patients with cancer and neutropenia: observations from two decades of epidemiological and clinical trials. Clin Infect Dis. 1997;25:247–259. [PubMed]
5. Ferraro M J, Gilligan P H, Saubolle M A, Weissfeld A S. Section V, bacteriology. In: Murray P R, Baron E J, Pfaller M A, Tenover F C, Yolken R H, editors. Manual of clinical microbiology. 7th ed. Washington, D.C.: ASM Press; 1995. pp. 246–662.
6. Fredricks D N, Relman D A. Improved amplification of microbial DNA from blood cultures by removal of the PCR inhibitor sodium polyanetholesulfonate. J Clin Microbiol. 1998;36:2810–2816. [PMC free article] [PubMed]
7. Goh S H, Potter S, Wood J O, Hemmingsen S M, Reynolds R P, Chow A W. HSP60 gene sequences as universal targets for microbial species identification: studies with coagulase-negative staphylococci. J Clin Microbiol. 1996;34:818–823. [PMC free article] [PubMed]
8. Greisen K, Loeffelholz M, Purohit A, Leong D. PCR primers and probes for the 16S rRNA gene of most species of pathogenic bacteria, including bacteria found in cerebrospinal fluid. J Clin Microbiol. 1994;32:335–351. [PMC free article] [PubMed]
9. Hayashi K, Yandell D W. How sensitive is PCR-SSCP? Hum Mutat. 1993;2:338–346. [PubMed]
10. Kloos W E, George C G. Identification of Staphylococcus species and subspecies with the MicroScan Pos ID and Rapid ID panel systems. J Clin Microbiol. 1991;29:738–744. [PMC free article] [PubMed]
11. Kullen M J, Brady L J, O'Sullivan D J. Evaluation of using a short region of the recA gene for rapid and sensitive speciation of dominant bifidobacteria in the human large intestine. FEMS Microbiol Lett. 1997;154:377–383. [PubMed]
12. Ley B E, Clinton C J, Bennett D M C, Jalal H, Foot A B M, Millar M R. Detection of bacteraemia in patients with fever and neutropenia using 16S rRNA gene amplification by polymerase chain reaction. Eur J Clin Microbiol Infect Dis. 1998;17:247–253. [PubMed]
13. Mermel L A, Maki D G. Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med. 1993;119:270–272. [PubMed]
14. Nucci M, Spector N, Bueno A, Solza C, Perecmanis T, Bacha P C, Pulcheri W. Risk factors and attributable mortality associated with superinfections in neutropenic patients with cancer. Clin Infect Dis. 1997;24:575–579. [PubMed]
15. Orita M, Suzuki Y, Sekiya T, Hayashi K. Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics. 1989;5:874–879. [PubMed]
16. Pittet D, Li N, Woolson R F, Wenzel R P. Microbiological factors influencing the outcome of nosocomial bloodstream infections: a 6-year validated, population-based model. Clin Infect Dis. 1997;24:1068–1078. [PubMed]
17. Pittet D, Tarara D, Wenzel R P. Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs, and attributable mortality. JAMA. 1994;271:1598–1601. [PubMed]
18. Reimer L G, Wilson M L, Weinstein M P. Update on detection of bacteremia and fungemia. Clin Microbiol Rev. 1997;10:444–465. [PMC free article] [PubMed]
19. Relman D A, Schmidt T M, MacDermott R P, Falkow S. Identification of the uncultured bacillus of Whipple's disease. N Engl J Med. 1992;327:293–301. [PubMed]
20. Thorpe T C, Wilson M L, Turner J E, DiGuiseppi J L, Willert M, Mirrett S, Reller L B. BacT/Alert: an automated colorimetric microbial growth detection system. J Clin Microbiol. 1990;28:1608–1612. [PMC free article] [PubMed]
21. Tumbarello M, Tacconelli E, Donati K G, Leone F, Morace G, Cauda R, Ortona L. Nosocomial bloodstream infections in HIV-infected patients: attributable mortality and extension of hospital stay. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;19:490–497. [PubMed]
22. Weinstein M P, Towns M L, Quartey S M, Mirrett S, Reimer L G, Parmigiani G, Reller L B. The clinical significance of positive blood cultures in the 1990s: a prospective comprehensive evaluation of the microbiology, epidemiology, and outcome of bacteremia and fungemia in adults. Clin Infect Dis. 1997;24:584–602. [PubMed]
23. Widjojoatmodjo M N, Fluit A C, Verhoef J. Molecular identification of bacteria by fluorescence-based PCR-single-strand conformation polymorphism analysis of the 16S rRNA gene. J Clin Microbiol. 1995;33:2601–2606. [PMC free article] [PubMed]
24. Woese C. Bacterial evolution. Microbiol Rev. 1987;51:221–271. [PMC free article] [PubMed]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links