Testing for antibiotic susceptibility

In most cases of infections, clinicians prescribe a course of antibiotics on the basis of a patient's symptoms and medical guidelines. In many situations, however, the bacteria will be resistant to several classes of antibiotics and local information is lacking, so it is critical to carry out an antimicrobial susceptibility test (AST). Performed routinely in clinical laboratories, the AST helps to confirm which antimicrobial drug will successfully inhibit the growth of the bacteria or another microbe causing the infection in a patient.

Susceptibility testing is not only important for helping to select the right treatment for individual patients. It also provides a tool for detecting resistance at the population level, which is key to monitoring the incidence and prevalence of AMR. Thus, AST can build up that vital local knowledge. On both fronts, AST is an important tool for guiding the appropriate use of antimicrobial drugs and preserving their efficacy.

One of the first AST methods was developed by Alexander Fleming. In the early 1920s he devised a technique to help him investigate the antibacterial effect of lysozyme, a naturally occurring enzyme found in tears, saliva and milk. It involved cutting a simple ditch in an agar plate and then filling it with the lysozyme-containing substance. Once the agar solidified and the lysozyme containing fluid was absorbed, he then streaked different microbes from the furrow to the edge of the plate. After a few hours of incubation he inspected the plate to see which microbes had grown or been inhibited by the antibacterial substance. Fleming used the same technique, with the addition of broth to the agar mixture, to assess the antibacterial impact of different concentrations of penicillin which he found spread rapidly through the agar. Based on his agar diffusion test, Fleming was able to work out which microbes were sensitive to penicillin, and to determine the minimum concentration needed to achieve inhibition (Fleming, 1929; Fleming, 1945).

The use of an agar diffusion method to study the impact of different substances on bacteria was not new. As early as 1889 Martinus W Beijerinck, a Dutch microbiologist and botanist based at the Netherlands Yeast and Alcohol Factory at Delft, described a similar technique to investigate the effect of different plant hormones on the growth of bacteria. Fleming and others continued to refine the agar diffusion technique, which became an important tool in the production of penicillin. A number of new methods had been introduced by the 1940s, which moved the process away from putting antimicrobial substances into a ditch cut in agar. Many of the changes were made because the original agar diffusion ditch method was tricky, labour intensive and time consuming (Heatley; Vincent, Vincent; Wheat).

One the new techniques, described in 1941, was developed by the British biochemist Edward Abraham and his colleagues at Oxford University. They placed open-ended cylinders on the surface of the agar plate seeded with the test organisms filled with an antimicrobial substance. Another involved placing absorbent paper discs impregnated with the antimicrobial solutions on the agar plate. First proposed in 1940 by CG Pope from the Wellcome Physiological Research Laboratories in England (Heatley), filter paper discs soon proved highly popular, with many scientists deploying them for AST by the late 1940s. Tablets containing penicillin were also used for the same purpose (Wheat, Spencer).

Another early dilution method was one that was performed in a test tube. Devised in the early 1870s, the method facilitates the growth and identification of bacterial populations suspended in a liquid diluted medium. The most commonly used mediums are agar and broth. In 1929 Fleming outlined a serial dilution technique to understand the activity of antibiotics. His method involved mixing a two-fold dilution of penicillin with a pre-inoculated liquid medium and checking the degree to which the suspension became cloudy to determine the antibiotic's activity. He later modified the protocol in 1942 which involved measuring the pH of the suspension instead of checking how cloudy it was.

The year 1942 also saw Charles Rammelkamp and Thelma Maxon introduce a broth marco dilution technique, also known as the 'tube dilution method'. Their method involved exposing bacteria to decreasing concentrations of antimicrobial agents in liquid media by serial two-fold dilutions. The test made it possible to establish in vitro the antimicrobial susceptibility of bacterial isolates from clinical specimens. It also provided a tool to work out how far the concentration of an antimicrobial could be lowered before it no longer inhibited the growth of bacteria. Known as the minimal inhibitory concentration (MIC), this was determined by watching out for when there was no visible bacterial growth (Khan, Siddiqui, Park).

Having a means to test for susceptibility was particularly important in the early years when the production of penicillin remained limited and extremely expensive so needed to be reserved for patients with the most need (Amsterdam). Such testing continued to be just as essential with the arrival of new antibiotics and by the early 1950s the paper disc diffusion method had become a common procedure in most clinical microbiology laboratories (Hudzicki).

One of the major drawbacks with the early dilution methods was they required a large volume of reagents and many tedious steps. There was also a strong possibility of false positive results due to long incubation times and cross-contamination. On top of this, many laboratories modified the process to suit their own purposes, such as changing the incubation times, temperature, media and pH parameters. This made it difficult to compare the results of one laboratory with another, which hindered the effort to track the frequency and extent of resistance in bacterial species in different institutions and countries.

The diagnostic landscape had become so confusing by the late 1950s that it was clear that the process was in urgent need of standardisation. To this end, in July 1960, the WHO set up an expert committee to oversee the process, focusing in particular on disc diffusion tests (WHO; Wooton, MacGowan, Howe). In 1966 the WHO selected the Kirby-Bauer disc diffusion test to become the gold standard for antimicrobial susceptibility testing. It was a modified version of a test first developed in 1956 by W. M. M. Kirby and A Bauer at the University of Washington School of Medicine and the King County (Bauer et al).

The Kirby-Bauer method had the advantage that it was simple and cost effective which made it suitable for routine testing in clinical microbiology laboratories. It also facilitated the easy testing of multiple antibiotics against the same target (Wooton, MacGowan, Howe). Many other standardised disc diffusion tests followed, with guidelines issued by organisations, such as the US National Committee for Clinical Laboratory Standards, and groups like the Swedish Reference Group and Deutsches Institut fur Normung, to iron out testing irregularities between laboratories and countries. The guidelines concentrated on defining the performance limits for growth medium, incubation conditions, inoculum concentrations, disc content, inhibition zone diameters, interpretative criteria and the parameters for quality control (Felmingham, Brown).

For a long time, AST remained a largely manual process. One of the first mechanisations was introduced to the serial dilution AST. In 1956 a microtitration system, using calibration spiral wire loops and droppers, was introduced. This improved the speed and accuracy of the dilution procedure. Another instrument, called the Multiple Inoculator was devised in 1959 which helped to seed bacteria on the agar plate for the routine testing of bacteria against antibiotics. This, however, only helped with one of the multiple steps needed to carry out an AST. By the late 1960s the momentum for automation had begun to pick up because of the great increase in the volume of specimens that needed to be processed by clinical laboratories and advances in the interface between laboratories and hospital information systems.

The first automated system, the Autobac disc elution system, was introduced by Pfizer Diagnostics in 1974. It used angle light scattering to compare microbial growth in two different broths. The first served and a control broth and the second contained antimicrobial drugs dispensed by discs in a disposable plastic multi-chambered cuvette. The machine came with an electronic data processing package that linked the results with a central computer to provide antibiotic susceptibility patient profile data (Gall). Able to provide results within 4 to 6 hours of inoculation, the machine was designed to test Enterobacteriaceae, Pseudomonas spp., staphylococci and enterococci. It could test up to 12 drugs at a time (Piddock). By 1982, the machine had been expanded to allow for the rapid testing of 30 different groups of gram-negative bacilli and for urine screening, and had reduced the incubation time needed to 3 hours (Sielaff, Matsen, McKie).

A number of other automated systems followed the Pfizer machine. In 1977 Abbott Diagnostics introduced a similar system, the MS-2 system, and McDonnell Douglas Corporation began marketing the AutoMicrobic System (AMS). The AMS was developed as part of McDonell's programme to detect and identify microorganisms found in space. It included a disposable miniaturised plastic specimen-handling system, solid-state optics for microbial detection and a mini-computer (O'Hara). The AMS was a predecessor of what is now known as the Vitek System, introduced in the early 1980s. Able to provide results in 4 hours, the Vitek System deployed dehydrated reagents in sealed plastic cards and contained separate cards for AST and organism identification (Wheat).

Despite the automation, susceptibility testing has continued to remain a slow process. Part of the problem is that pathogens need to be isolated and grown from clinical specimens before AST can be performed. Depending on which microbe is involved, this process can be very slow. On average it takes 2 to 3 days to get results from bacterial cultures, but in some cases it can take much longer. Mycobacteria that cause tuberculosis, for example, can take 30 days to culture. Faced with a patient with a severe infection, doctors are therefore usually forced to prescribe an antibiotic without knowing the AST result (Bergeron, Ouellette).

References

Amsterdam, D (1988), 'Principles of antibiotic testing in the laboratory', in A Balows et al (eds), Laboratory Diagnosis of Infectious Diseases, chapter 4.Back

Bauer, AW, Kirby, WMM, Sherris, JC, Turck, M (1966), 'Antibiotic susceptibility testing by a standardized single disk method', American Journal of Clinical Pathology, 45/6, 493-96.Back

Bergeron, MG, Ouellette, M (Aug 1998), 'Preventing antibiotic resistance through rapid genotypic identification of bacteria and of their antibiotic resistance genes in the clinical microbiology laboratory', Journal of Clinical Microbiology, 36/8, 2169-72.Back

Heatley, NG (18 Dec 1944), 'A method for the assay of penicillin', Biochemical Journal, 38/1, 61-65.Back

Fleming, A (1929), 'On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B influenzae', British Journal of Experimental Pathology, 10, 226-36.Back

Fleming, A (11 Dec 1945) 'Penicillin', Nobel Lecture.Back

Felmingham, D, Brown, DFJ (2001) 'Instrumentation in antimicrobial susceptibility testing', Journal of Antimicrobial Chemotherapy, 48, Supplement S1, 81-85.Back

Foster, JW, Woodruff, HB (10 Nov 1943), 'Microbial aspects of penicillin. VI Procedure for the cup assay for penicillin', Journal of Bacteriology, 46/5, 43-58.Back

Friedman, MacLowry, J (Sept 1973), 'Computer identification of bacteria on the basis of their antibiotic susceptibility patterns', Applied Microbiology, 36/3, 314-17.Back

Gall, LS, Curby, WA (2018) Instrumental Systems for Microbiological Analysis of Body Fluid.Back

Hudzicki, J (8 Dec 2009) 'Kirby-Bauer Disk Diffusion Susceptibility Test Protocol', American Society for Microbiology, 58.Back

Khan, ZA, Siddiqui, MF, Park, S (2019) 'Current and emerging methods of antibiotic susceptibility testing', Diagnostics.Back

O'Hara, CM (Jan 2005) 'Manual and automated instrumentation of identification of Enterobacteriaceae and other aerobic Gram-Negative bacilli', Clinical Microbiology Reviews, 18/1, 147-62.Back

Piddock, LJV (1990) 'Techniques used for the determination of antimicrobial resistance and sensitivity in bacteria', Journal of Applied Bacteriology, 68, 307-18.Back

Scholthof, K BG (2001) '1898 - the beginning of virology … time marches on'.Back

Sielaff, BH, Matsen, JM, McKie (June 1982) 'Novel approach to bacterial identification that uses the Autobac System', Journal of Clinical Microbiology,, 15/6, 1103-10.Back

Thornsberry, T (1985), 'Antimicrobial Susceptibility Tests: Automation and Mechanization', in Habermehl KO, eds, Rapid Methods and Automation in Microbiology and Immunology.Back

Vincent JG, Vincent HW (1 March 1944) 'Modification Oxford cup penicillin determination', Experimental Biology and Medicine.Back

Wheat, PF (2001) 'History and development of antimicrobial susceptibility testing Methodology', Journal of Antimicrobial Chemotherapy, 48, Suppl. S1, 1–4.Back

Wheat PF, Spencer, RC (5 Nov 1988), 'The evolution of in-vitro antimicrobial susceptibility techniques', Journal of Antimicrobial Chemotherapy, 22/5, 579-82.Back

WHO (11-16 July, 1960) Standardization of Methods for conducting Microbic Sensitivity Tests. Second Report of the Expert Committee on Antibiotics.Back

Wooton, M, MacGowan, AP, Howe, RA (2016) 'Towards better antimicrobial susceptibility testing', Journal of Antimicrobial Chemotherapy, 72, 323-29.Back