Laboratory automation impact on antimicrobial resistance

Antibiotic resistance in common bacterial pathogens, such as Staphylococcus aureus, Escherichia coli, and Klebsiella pneumoniae, has significantly limited the therapeutic options available for management of infectious diseases. While the indiscriminant use of broad spectrum antibiotics is a significant contributing factor, a more fundamental problem exists. Diagnostic microbiology test results have historically been available too late to be useful. This is, in part, due to the nature of the test methods and in part due to workflow practices in the laboratory. Thus, patients remain on empiric treatments that are frequently ineffective or unnecessarily too broad spectrum^1,2. Microscopy and bacterial cultures are mainstays in the microbiology lab, using techniques developed more than 100 years ago. Although microbiologists speak with pride about the ‘art’ of their science, the clinical value of the diagnostic tests is frequently lost because of the delays in reporting results with these ‘traditional’ approaches. Fortunately, the practice of clinical microbiology is undergoing a dramatic transformation with the introduction of molecular diagnostics, primarily for rapid diagnosis of infections caused by viruses and difficult to grow bacteria, MALDI-TOF mass spectrometry for identification of bacteria, mycobacteria and fungi, and automation of all practices in bacteriology.

Whereas the impacts of the molecular diagnostics and identification by mass spectrometry on time to results are readily recognised, the transformative impact of automation is only slowly being realised, in part because the full benefits of automation require the laboratory to change their work practices. Processing specimens in the microbiology laboratory can be subdivided into: (1) the initial receipt and processing of the specimen (e.g. preparation of Gram stain and inoculation of plates); (2) the incubation of the inoculated plates; (3) examination of the plates; and (4) performance of identification tests and antibiotic susceptibility tests. The impact of lab automation will have on improving patient outcomes and combatting antimicrobial resistance is directly related to improving diagnostic accuracy and timeliness of results.

The traditional approach after receipt of a specimen is to enter the patient demographic information and test requests into the laboratory information system, sort the specimens based on processing priorities, and then inoculate specific enrichment and selective differential culture media based on established laboratory procedures. Because the selection of culture media is defined by the specimen type, the common practice is to process similar specimens in batches. This is efficient for the technologist but less so for processing specimens with a lower priority – processing these specimens may introduce significant delays from the time of receipt to inoculation onto media and subsequent incubation. Additionally, because the processing area and incubators are typically in different areas of the laboratory, it is inefficient for the technologist performing the initial processing to transport the plates to the incubator after each specimen is processed. This is normally done in batches or even at fixed intervals during the day, introducing significant delays between inoculation of plates and incubation. Finally, the inoculation of plates is dependent on the skill of the technologist. Indeed, the traditional approach to inoculation of a plate is to perform a four-quadrant streaking pattern (with ideally the use of a sterile loop for each quadrant) in order to obtain adequate isolation of colonies for further workup. The quality of this process can vary tremendously based on the technical skills of the microbiologists.

The first automated systems focused on the initial processing of specimens because it was recognised that the quality of this step was important for all subsequent work. Full laboratory automation (Figure 1) allows the user to process specimens immediately on receipt in the laboratory and then automatically transport the inoculated plates to incubators by a track system similar to what is used in automated chemistry, haematology, or immunology laboratories. Systems that permit inoculation of multiple plates simultaneously result in more rapid throughput than possible with inoculation of single plates manually. Use of a pipette to dispense the sample rather than the traditional bacteriologic loop ensures the proper volume is dispensed, and the use of magnetic beads or mechanical loops improves both the reproducibility of the plate inoculation and the ability to obtain isolated colonies^3–6. Improved specimen throughput, rapid incubation and decreased need for subcultures all directly impact on the time from initial specimen processing to identification and antimicrobial susceptibility testing, critical steps in moving from empiric antibiotic therapy to directed therapy.

Figure 1.  BD Kiestra total laboratory automation platform.

Traditionally, inoculated plates are incubated overnight before they are examined. Because plates are inoculated throughout the day and evening, they have been typically incubated 12–24 hours before they are removed from the incubator. Laboratory automation systems allow standardisation of the incubation time for each specimen type (i.e. shorter incubation times for specimens such as urines with rapidly growing organisms, longer incubation times for specimens such as respiratory with slow growing organisms)^7–10. In other words, incubation is defined by the specimen type and not by the work practices of the technical staff. Indeed, this is a challenge for the laboratory to maximise the value of automation by processing the specimen after it has been incubated for a defined period of time which can occur throughout the day, evening and night.

The common practice in manual processing of cultures is to remove all culture plates when the technical staff arrives in the laboratory, sort the plates by specimen type, and then process the cultures by technologists working with the specific specimens (e.g. urines, respiratory, wounds, stools, etc.). There are some efficiencies realised by having a technologist work with the same specimen type, with each culture processed according to standardised procedures, and the same pathogens commonly isolated. However, the workload is unevenly distributed so it is difficult to assess the efficiency of individual technologists. Additionally, because the culture plates are not in the lab incubator during the processing, cultures may remain at room temperature for hours thus negating the advantage of incubation in the optimum temperature and atmosphere. With laboratory automation systems, plates remain in the incubator and examination is performed digitally by imaging the plates. Software has been developed to maximise the early detection of growth and optimise the images. Under the optimised incubation conditions, cultures can be examined and subsequent processing (e.g. identification, antimicrobial susceptibility tests) initiated hours earlier than traditional approaches^9,11,12. Better recovery of organisms such as S. aureus and E. coli as well as slow growing pathogens has also been reported with these incubation systems^13,14. The imaging software can also screen cultures for insignificant growth and eliminate further processing of the culture plates. Again, the decreased time to results and improved recovery directly impact on the timely selection of optimum therapy.

The use of MALDI-TOF mass spectrometry has dramatically changed the way organisms are identified, providing increased accuracy, decreased time to results and decreased costs^15–19, particularly when combined with total lab automation^9,11. Use of molecular diagnostics permits the rapid detection of resistance genes in selected pathogens; however, the presence of a resistance gene informs which drug cannot be used but not which drug can be used. Thus, guidance for the selection of antibiotic treatment currently requires performance of traditional, phenotypic antibiotic susceptibility tests. Traditional susceptibility tests frequently require subcultures to obtain isolated colonies for testing. Automated platforms have been demonstrated to provide improved isolation of colonies thus eliminating the need for a significant number of subcultures, as well as reducing problems with mixed cultures requiring repeat testing.

Because empiric antibiotic therapy is frequently inappropriate, the rapid transition to directed therapy defined by the microbiological culture results is critically important for improved patient outcomes, decreased hospital costs and decreased risk of selecting antimicrobial resistant bacteria. The microbiology laboratory is undergoing a technological transformation with laboratory automation potentially influencing every stage of generating timely test results.

The author is a salaried employee of Becton Dickinson Diagnostic Systems.