Antimicrobial Chemotherapy, 5th Edition
Part 3 - General Principles of Usage of Antimicrobial Agents
Use of the Laboratory
This chapter covers the basic principles and limitations of using a microbiology laboratory to obtain information on the selection and control of antimicrobial therapy. It is not a comprehensive account of the clinical laboratory microbiology. The examples used refer primarily to bacteriological practice, but the principles apply to the investigation of any infection.
In the diagnostic microbiology laboratory the aim is to identify the presence of pathogens as rapidly as possible and to provide, where applicable, antimicrobial susceptibility data to the clinician. A wide range of techniques, including microscopy, culture, antigen or antibody detection, and nucleic acid detection methods are used in the diagnostic laboratory. There is an increasing trend to standardize the protocols used to detect micro-organisms and their antimicrobial susceptibility. When using laboratory services, it is important to provide appropriate clinical details (e.g. travel history, antibiotic therapy) on the request form; without these the optimal use of diagnostic methods cannot be guaranteed. Clinical details may dictate the tests used and influence the interpretation of the result; e.g. in a patient treated with gentamicin (and a penicillin) for endocarditis the optimal serum aminoglycoside concentration is lower than that required in other infections. Crucially, many diagnostic methods, in particular those involving microbial culture are prone to biological variability, which may hinder the interpretation of results.
Clinical laboratories rely on the quality of the specimens they receive; none more so than microbiology departments where the final result may depend on the degree of care observed in taking the specimen. A single contaminating bacterium introduced into a blood culture during collection may result in a patient being incorrectly diagnosed as having bacteraemia. Similarly, extraneous nucleic acid contaminating a sample can cause a false positive result, for example in mid-stream samples from women tested for chlamydial or gonococcal infection. Such an error could have profound consequences.
Some of the more common problems are listed.
- Inappropriate specimen.Saliva is submitted instead of sputum; a superficial skin swab is taken instead of a swab of pus (a specimen of pus in a sterile bottle is always preferable to a swab when possible).
- Inadequate specimen.The specimen may be too small (especially fluids for culture for tubercle bacilli). Rectal swabs are not an adequate substitute for faeces. It is almost impossible to interpret the significance of microbes cultured in tracheal aspirate in a patient who is ventilated, because of the likely presence of colonizing upper respiratory tract flora. Deep respiratory specimens, such as broncho-alveolar lavage fluid, are highly preferable to make an accurate diagnosis of pneumonia in such critically ill patients.
- Wrong timing.Specimens taken after the start of chemotherapy, when the causative organism may no longer be demonstrable, or after the patient has recovered. It is not uncommon for the laboratory to receive rock-hard faeces from patients with ‘diarrhoea’. Many laboratories will not process non-diarrhoeal faecal samples.
- Wrong container.Blood for culture put in a plain (sometimes non-sterile) bottle instead of the correct culture fluid; biopsies put into bactericidal fixatives such as formalin.
- Clerical errors.Incorrect labelling; incomplete or misleading information on request forms.
Prompt specimen transport to the laboratory is essential. Material submitted for culture may contain living cells; any delay in reaching the optimal cultural conditions will result in loss of viability. With fastidious organisms such as gonococci or viruses this may result in failure to isolate the organism. Conversely, overgrowth of pathogens by fast-growing commensals also commonly occurs during the period between collection of the specimen and processing in the laboratory, potentially obscuring true pathogens or resulting in a false positive result. For example, urine that is left at room temperature will act as a culture medium; bacteria that may be present in only low numbers may multiply to levels above the quantitative threshold that is used to define a positive result.
Specimens from potential medical emergencies, such as bacterial meningitis or malaria, should be delivered to the laboratory as soon as possible after collection for immediate processing. For swabs, most laboratories recommend a form of suspended animation in which the specimen is placed in a special transport medium comprising soft, buffered agar containing charcoal to inactivate any toxic substances. The effect of transport delays on microbe survival can be minimized by inoculating culture media next to the patient and incubating them immediately. This may be achieved in special units with laboratories attached (for example, in some genito-urinary medicine clinics), but is not practicable in most situations. An exception is blood culture, where the counsel of perfection should apply. Near patient tests are becoming more widely available, for example the detection of Group A streptococcal infection in patients presenting with pharyngitis. Unfortunately, however, such rapid detection methods do not permit the assessment of the antimicrobial susceptibility of the pathogen, meaning that conventional culture may be required as a supplementary test.
The flow diagram (Fig. 12.1) outlines the main steps that occur when a specimen is submitted for bacteriological investigation: microscopy (especially for specimens from normally sterile sites), culture, identification, and antimicrobial susceptibility testing. Even with the most rapidly growing bacteria and with improved methods, results of culture and sensitivity often take 48 h. This may be further delayed if there is a mixture of organisms or if slow-growing pathogens, such as Mycobacterium tuberculosis, obligate intracellar pathogens (e.g. chlamydiae) or viruses, are involved. Not all micro-organisms are readily cultivable and a report of ‘sterile’ or ‘no growth’ does not definitively mean that the specimen contained no organisms, but rather that the laboratory was unable to isolate a pathogen from the specimen. Conversely, many diagnostic specimens will grow microbes that would normally be expected to be found in particular anatomical sites, and it is important to distinguish these from pathogens; in practice this may be hard or impossible to achieve with certainty. A prime role of the medical microbiologist is to interpret such results, and advise on further testing and treatment as appropriate.
Because of the inevitable delay in obtaining culture results there is a need to inform clinicians of important preliminary findings before the complete results are known. Microscopy results are usually available on the same day as the specimen is received and in urgent cases can be reported within 1 h or less. For example, a Gram-film of cerebrospinal fluid can be done very quickly and may give the physician a reliable guide to primary therapy (which may be life saving) while waiting for cultural confirmation of the result. Similarly, positive blood cultures and other findings serious to the individual patient or his immediate contacts are usually telephoned directly to the doctor. When the antibiotic sensitivity is predictable (e.g. Streptococcus pyogenes is always sensitive to penicillin) this advice may be given with the initial report. With many bacteria the report ‘susceptibility to follow’ is all that can be imparted before antimicrobial testing, although a ‘best guess’ based on known patterns of resistance in the hospital or community may be suggested.
Fig. 12.1 The various steps between obtaining a specimen from a patient and the issue of the final report. Note the importance of the period before the specimen arrives in the laboratory: unless the specimen is properly taken and transported it may be useless, and unless the request card is properly completed wrong tests may be done.
Although many diseases have a well-defined microbial aetiology, the information obtained about patients and the results of specimen examination are often too sparse to form a definitive opinion as to the microbial cause in an individual case. Most patients with infection survive and many improve so rapidly that the significance of microbes isolated is never known. Thus, although we know from historical and epidemiological evidence that Str. pyogenes causes tonsillitis and is involved in the aetiology of rheumatic fever, when an individual patient presents with joint pains following a sore throat we cannot be absolutely sure that the Str. pyogenes in the throat is the cause of the illness, as opposed to representing the chance finding of asymptomatic colonization in someone with disease due to another cause. In such situations, additional information (such as raised antistreptolysin O antibodies in the serum in this case) may be needed in order to establish a causal relationship.
The greatest difficulty is encountered with specimens from areas of the body that have a resident microbial flora that may sometimes assume a pathogenic role. Isolations from respiratory sources are often the most difficult to interpret. Demonstration of M. tuberculosis is always significant, and even in the absence of clinical disease the patient must be further examined and treated. Conversely, Candida albicans is often a harmless commensal, although it can cause serious disease in immunocompromised patients. The ability of some organisms to strike while the host defences are down has given them the title opportunist pathogens. However, merely by looking at cultures of these opportunists, it is impossible to say in a particular case whether or not they are adopting a pathogenic role. The interpretation of their detection can be influenced by ancillary findings (e.g. presence or absence of pus; numbers of organisms isolated) and importantly by clinical information provided on the request card. If the latter is absent, non-contributory, or misleading, as it frequently is, the report issued may be valueless.
In the absence of sufficient supportive information, the laboratory can adopt one of two approaches: report any microbe isolated regardless of any possible significance, or report only common pathogens and dismiss all the others as ‘normal flora’. The importance of this to antimicrobial therapy is that if the isolate is not considered ‘significant’, no further work (including antimicrobial susceptibility tests) will be carried out. The corollary is that many susceptibility tests may be carried out, some unnecessarily. This occurs more commonly than is usually admitted. In turn patients may be prescribed unnecessary and potentially toxic antimicrobial drugs because commensal bacteria isolated from a badly taken specimen were considered significant and susceptibility results issued. Sometimes a great deal of effort and expense is put into treating colonizing organisms that are merely filling a vacuum left by the normal flora and would quietly disappear if antimicrobial chemotherapy were withheld.
Rapid advances in immunological and molecular techniques continue to provide new antigen and nucleic acid detection methods. These were initially used to make a diagnosis of infection where viruses or other difficult-to-culture pathogens were suspected. However, the diagnosis of some infections has been transformed by the availability of nucleic acid amplification tests. For example, detection of meningococcal-specific DNA in blood or cerebrospinal fluid makes specific diagnosis of a potentially life-threatening infection possible within hours. This approach can also identify the Neisseria meningitidis group and so detect early clusters of cases. One drawback of this approach is that without a viable pathogen it is not be possible to perform standard antimicrobial susceptibility tests (see below). Viral load measurements for the human immunodeficiency virus (HIV) are now routinely performed by a quantitative polymerase chain reaction method as part of HIV disease management. The results are used to assess patient prognosis and the effectiveness of antiretroviral therapy. Periodic monitoring of viral load can promptly identify treatment failure potentially due to the emergence of resistance to antiviral drugs.
Micro-array technology is a method of DNA analysis that involves fixing potentially thousands of DNA probes on to a glass slide. After exposing the slide to a specimen containing pathogens, the DNA fragments that have bound to the probes are detected by chemiluminescence or fluorescence systems, the sensitivity of which may be increased using the polymerase chain reaction. This approach has been successfully applied to the detection of rifampicin-resistant strains of M. tuberculosis, so allowing informed choices to be made at the start of therapy. Such approaches offer an alternative source of antimicrobial susceptibility information. However, when many different mechanisms of resistance are possible, some of which may not be characterized by the presence of specific genes, then susceptibility testing by these approaches may not be possible. It is also possible that some genes although present may not be expressed in vivo.
Techniques based on micro-organism genotype (e.g. DNA fingerprint) rather than phenotype (e.g. whole cell protein and lipopolysaccharide profiles, antibiotic susceptibility profile, biochemical tests) are now used preferentially to determine the relatedness of clinical isolates. These techniques are useful in the investigation of outbreaks of infection, and in determining routes and sources of infection, including clusters of cases caused by antibiotic-resistant pathogens.
Antimicrobial susceptibility testing
Purpose of susceptibility testing
Since therapy of infection normally begins, quite properly, before laboratory results are available, antibiotic susceptibility testing primarily plays a supplementary role in confirming that the organism is sensitive to the agent that is being used. Sometimes it may enable the prescriber to change from a toxic to a less toxic agent, or from an expensive to a cheaper one.
Usually the laboratory report will influence treatment only if the patient is failing to respond. By this time, the laboratory should have succeeded in establishing the susceptibility pattern of the offending organism (if it is bacterial) and can advise the clinician as to how treatment might be modified. Susceptibility testing of non-bacterial pathogens is not usually possible, although limited antifungal testing is carried out in some centres. Antiviral susceptibility testing is becoming available for certain viruses, for example HIV, cytomegalovirus, and herpes simplex; these tests require isolation of the virus and results are not usually available for a week or more.
Many laboratories record and disseminate data on the susceptibility patterns of common pathogens in the hospital and in the community to aid the choice of effective therapy. Patterns of bacterial susceptibility and resistance vary considerably from place to place and hospitals, or even wards, often have their own particular resistance problems so that the results need to be tailored to the individual circumstance. Regional, national, and international resistance trends are monitored by laboratory networks reporting to a central point.
Most diagnostic microbiology laboratories test antibiotic susceptibility of bacteria by some form of agar diffusion test in which the organism under investigation is exposed to a diffusion gradient of antibiotic provided by an impregnated disc of filter paper. Some use one of the semi-automated commercial devices that provide susceptibility estimates with various degrees of sophistication. The disc diffusion method is flexible, simple, and cheap and a result can usually be obtained within a day with rapidly growing pathogens such as Staphylococcus aureus, Pseudomonas aeruginosa, and various enterobacteria; it is less suitable for fastidious or slow-growing bacteria such as anaerobes, streptococci, Haemophilus spp., and Neisseria spp., for which alternative procedures are preferable. Mycobacterium tuberculosis is very slow growing and needs special media for cultivation and susceptibility testing. Molecular methods that are able to detect DNA sequences associated with resistance traits (see above) are gradually becoming more common.
Minimum inhibitory and bactericidal concentrations
A more accurate estimate of the susceptibility of a bacterial isolate to antimicrobial agents can be obtained by titration in broth or on agar plates containing graded dilutions of antibiotics. The concentration that completely inhibits growth after a defined incubation period (usually overnight) is known as the minimum inhibitory concentration. The minimum inhibitory concentration can also be more easily (but more expensively) estimated by use of a commercial variant of the disc diffusion test, the ‘Etest’.
If broth dilution procedures are used the minimum bactericidal concentration of antibiotic for the test strain may additionally be determined if so desired. The criterion for bactericidal activity is generally taken to be a 1000-fold reduction in the original inoculum after overnight incubation. This end-point is entirely arbitrary and takes no account of the rate of killing, which may be more important.
Estimation of bactericidal concentrations achieved in the patient's serum during therapy is sometimes used in infections (notably bacterial endocarditis) in which bactericidal activity is essential to a cure. The patient's serum, obtained 1 h after a dose, is titrated against the organism responsible for the infection (so-called ‘back titration’). The results are susceptible to methodological variation and their interpretation has been widely questioned; indeed recent guidelines do not recommend the use of these tests.
Clinical relevance of antibiotic sensitivity tests
A laboratory report of susceptibility or resistance by no means guarantees that the results will translate into clinical success or failure if the agent is used in therapy. Patients may fail to respond to antibiotics judged to be fully active against the offending microbe, or may recover despite the use of agents to which the organism is resistant. These situations arise because laboratory tests offer relatively crude estimates of susceptibility that fail to take into account many crucial features of the infection in the patient (Table 12.1). None the less, antibiotic sensitivity testing offers a generally reliable guide to therapy, particularly in the seriously ill patient in whom laboratory tests may provide an indispensable guide to patient care.
What the laboratory reports
The final report reaching the clinician must be self-explanatory, even dogmatic. It is not practicable, or desirable, to test each organism isolated against all antibiotics. A restricted range of antimicrobial agents is usually tested against isolates considered significant, with a different selection for Gram-positive and Gram-negative bacteria. Primary testing is ordinarily restricted to a few old and well-tried agents that are perfectly adequate for most common infections (Table 12.2). More extensive (second-line) testing, particularly of expensive, broad-spectrum agents, is reserved for resistant isolates or bacteria from patients with serious infections that are presenting problems of management. Usually, only two or three of the susceptibilities tested are reported even if more are performed. Such restricted reporting has the important function of reinforcing local antibiotic policies and of discouraging clinicians from using inappropriate agents. Extended testing may also provide useful epidemiological information of trends of antimicrobial susceptibility and of clusters of multiply resistant bacteria, indicating cross-infection or spread from a common source.
Table 12.1 Some aspects of infection that may cause the results of in-vitro tests not to be reflected during treatment
Most infections are caused by a single organism, often a well-known pathogen. In these cases there is usually no problem in deciding what to test and report. However, some specimens (e.g. those from abdominal wounds) are often infected with mixtures of organisms and each must be individually identified and tested against appropriate antimicrobial agents. Laboratories usually restrict the reporting of such results to discourage a blanket therapy approach covering each and every micro-organism that is recovered from specimens taken from normally non-sterile sites.
The choice of antibiotic tested varies with the site of infection and the pharmacological properties of the drug. Some agents, such as nitrofurantoin and nalidixic acid, achieve therapeutic concentrations only in urine and are of no value in other infections. Information about the individual patient may alter drug testing:
- if the patient is allergic to penicillins alternatives will be sought
- in pregnancy, sulphonamides and trimethoprim should be avoided if possible because of the risk of folate deficiency
- in a patient with worsening renal failure it may be preferable to avoid aminoglycoside therapy
- tetracyclines should not be used in late pregnancy or in young children owing to deposition in teeth.
These limitations of susceptibility testing or reporting can be taken into account by the laboratory only if the appropriate information is given on the request card.
Restrictions imposed by the large number of available antibiotics may be approached in various ways. Some groups of agents, such as aminopenicillins (ampicillin, amoxicillin) or tetracyclines, are so similar in terms of their antibacterial spectrum that only one representative of each needs to be tested. With other drugs where there is differential susceptibility of bacteria to different members of the group, such a decision is less easy. An organism susceptible to cefalexin, one of the earliest cephalosporins, is also usually susceptible to all subsequent members of that group and this is often used as a screen for cephalosporin-sensitive bacteria. However, the converse is not true; an organism resistant to cefalexin may be susceptible to later cephalosporins and a definitive statement in this regard can be made only by testing the appropriate compound. The same principle applies to nalidixic acid and newer, more active quinolones.
Table 12.2 Examples of a restricted range of antimicrobial agents selected for primary susceptibility testing of some common pathogens
Interpreting antimicrobial susceptibility reports
Consider a report on pus from an abscess that grew Staph. aureus, which was resistant to penicillin, but susceptible to erythromycin and cloxacillin. The statement ‘this organism is resistant to penicillin’ means that penicillin would not influence the outcome. The infection may well improve due to host defences or to adequate drainage of pus, but since penicillin-resistant staphylococci are invariably β-lactamase producers any penicillin that reached the abscess would be rapidly destroyed. Such a statement is based on sound laboratory and clinical evidence. If the Staph. aureus isolate was found to have relatively low-level resistance (i.e. where inhibition of growth is incomplete but less than that produced using a susceptible control organism) the likelihood of clinical resistance to penicillin is more difficult to determine; factors such as site of infection and drug penetration may be important in this respect. The laboratory will try to weigh up the evidence and score the result as ‘susceptible’ or ‘resistant’; the term ‘reduced (or intermediate) susceptibility’ is sometimes used, but this is a less satisfactory alternative and leaves the clinician uncertain how to interpret the result.
The statement ‘this organism is susceptible to cloxacillin’ implies that use of this antibiotic (or a related β-lactamase-stable penicillin) would influence the outcome. This is more difficult to support than a statement about resistance. Treatment with the antibiotic may elicit little response in the patient because insufficient drug may have penetrated into a large collection of pus; the dosage prescribed and route of administration may be important here. More importantly (although unlikely in the present example, since Staph. aureus is commonly incriminated in infected wounds) the wrong organism (an innocent bystander) may have been tested. Host factors that may also influence therapeutic outcome are described in Chapter 13.
When in doubt, whether about the optimal specimen to submit, the interpretation of a test result or the most appropriate treatment, the laboratory should be consulted. For unusual diseases and problem cases it is often possible to seek help from specialized units such as tropical hospitals and institutes. In some countries reference laboratories are available that provide expertise in particular areas. In the UK many of these operate under the aegis of the Health Protection Agency at Colindale. Worldwide, the Centers for Disease Control and Prevention, Atlanta, Georgia, USA, offer a service for the diagnosis and therapy of unusual infectious diseases.
Fawley WN, Wilcox MH. Molecular diagnostic techniques. Medicine 2005; 33: 26-32.
Yang S, Rothman RE. PCR-based diagnostics for infectious diseases: uses, limitations, and future applications in acute-care settings.Lancet Infectious Diseases 2004; 4: 337-348.