Diagnosis of infectious diseases
Newer DNA amplification methods have the potential significantly to influence the diagnosis and management of a variety of infectious diseases. Conventional laboratory diagnostic methods require a minimum of 24 hours and, in many cases, the time required is significantly longer. Moreover, cultures may yield no bacterial growth if there has been a delay in transporting the specimen to the laboratory, the number of viable infecting organisms is low, or the patient had been taking antibiotics at the time the culture specimen was obtained. In addition, certain pathogenic organisms, such as Mycoplasma sp, Chlamydia sp, rickettsia, and viruses, are not easily detected by routine culture methods and require specialized methods [32,33].
Rapid nonculture diagnostic tests relying on antigen detection by immunofluorescence or enzyme immunoassay, or using DNA probes, may have variable diagnostic sensitivities or specificities as compared with culture.
Molecular methods with amplification and detection of target nucleic acids generally have been found to have superior sensitivity and specificity and have the potential to provide results within hours of collecting the specimen [34-38]. Pilot studies have indicated the feasibility of designing broad-range multiplex PCR assays with the capability of detecting a panel of microorganisms from clinical specimens. Assays that are currently commercially available for use in diagnostic laboratories include tests for the detection of C trachomatis, Chlamydia pneumoniae, M tuberculosis, Mycoplasma pneumoniae (correlates with disease and establishes the diagnosis of atypical pneumonia), Neisseria gonorrhoeae, herpes simplex virus (HSV), cytomega-lovirus, enterovirus, and other infectious agents. In addition, there are PCR assays available for monitoring the viral load of HIV, hepatitis C virus, and hepatitis B virus. Unfortunately, only a few of these commercially available assays have been extensively evaluated to determine the sensitivity, specificity, or clinical usefulness. There are reliable commercial probe systems for detection of N gonorrhea, C trachomatis, and papilloma virus. PCR assays have been found to be significantly more accurate, with sensitivities of 90% to 100% and specificities greater than 97% for the detection of C trachomatis from cervical or urethral specimens. The positive predictive values reported in these studies ranged from 89% to 100%. A major advantage of these tests is the ability to detect Chlamydia in urine specimens or urethral swabs. PCR testing of freshly voided urine was found to be the most sensitive (91%) and most specific (100%) method for detecting asymptomatic C trachomatis infection in men . These assays also have been automated, allowing for the processing of large numbers of specimens. They may be used for diagnosis or screening for sexually transmitted diseases. A coampli-fication PCR assay for the direct detection of both N gonorrhoeae and C trachomatis from patients with sexually transmitted diseases has also been developed. The sensitivity and specificity of PCR detection of N gonorrhoeae from cervical and urethral specimens were found to be greater than 90% and 96%, respectively.
Direct amplification tests have also had a great impact on the rapid diagnosis of tuberculosis. Conventional culture methods for the isolation of my-cobacteria generally take several weeks, whereas PCR takes only 24 hours. Commercial amplification assays have been found to have sensitivities of about 90% to 98%, as compared with culture of specimens that are smear positive for acid-fast bacilli (AFB). The performance of these amplifications, however, has been suboptimal for specimens without AFB seen on direct microscopic examination, with reported sensitivities as low as 46%. The specificity of PCR-based assays for M tuberculosis is excellent at >98%, and sensitivity is at least 80%. Although these assays cannot replace myco-bacterial cultures, their ability rapidly to determine the presence of
M tuberculosis directly from respiratory tract specimens has enabled more rapid institution of effective therapy and implementation of important infection control and public health interventions [40-42]. Mycobacterium sp probes for the rapid identification of mycobacteria are now widely used to identify acid-fast organisms grown on solid media or in liquid media. Probes are available for M tuberculosis, M avium complex, M kansasii, and M gor-donae. Mixed infections (M tuberculosis and M avium-intracellulare) can be identified. At present, two Food and Drug Administration (FDA)-approved NAA assays are widely available: the AMPLICOR M. Tuberculosis (Roche Diagnostic Systems, Branchburg, New Jersey) and the Amplified Mycobacterium Tuberculosis Direct Test (Gen-Probe, San Diego, California). The AMPLICOR assay uses PCR to amplify nucleic acid targets, is FDA approved in smear-positive respiratory samples, sensitivity ranges from 74% to 92% for smear-positive samples and 40% to 73% for smear-negative samples, specificity goes from 93% to 99%, and it has a negative predictive value of 100%. The Mycobacterium Tuberculosis Direct Test assay is an isothermal strategy for detection of M tuberculosis rRNA and is FDA approved for smear-positive and smear-negative respiratory specimens. It has sensitivity from 83% to 98% from smear-positive respiratory samples and from 70% to 81% from smear-negative respiratory samples; it is very helpful for confirming disease in intermediate- and high-risk patients and for excluding cases in low-risk patients.
The Centers for Disease Control and Prevention now recommends that AFB smear and NAA be performed on the first sputum smear collected. If smear and NAA are positive, pulmonary tuberculosis is diagnosed with near certainty. If the smear is positive and the NAA is negative, is recommended to test the sputum for inhibitors by spiking the sputum sample with an aliquot of lysed M tuberculosis and repeating the assay. If inhibitors are not detected, the process needs to be repeated on another sample and if it still remains negative, the patient most likely has nontuberculous myco-bacterium. If smears are negative but clinical suspicion is high or intermediate, NAA should be done in a sputum sample, and if the sample test is positive a presumptive diagnosis of tuberculosis is made. If the smear is negative and the clinical suspicion is low NAA should not be done. Testing must be limited to those who have not been treated recently for active disease. The limitation of these NAA is that they give no drug susceptibility information; also, they detect nucleic acids from both living and dead organisms and may be false-positive for active disease. Assays that detect mRNA remain positive only while viable mycobacteria persist, so they are sensitive indicators of treatment response and drug susceptibility.
For extrapulmonary tuberculosis AFB smears and culture are less sensitive than in respiratory samples. Herein NAA play an important role in the diagnosis, although it is not completely defined. The sensitivity in nonrespir-atory samples for Mycobacterium Tuberculosis Direct Test goes from 67% to 100%; in smear-negative samples, the sensitivity ranges from 52% to
100% in different studies. The AMPLICOR had a similar sensitivity, and the specificity of both assays remains high in nonrespiratory samples. The assays perform different based on the type of sample (spinal fluid, pleural fluid, ascitic fluid). Multidrug resistance in tuberculosis is a major public health problem. Coupling assays that detect gene mutations, such as line probe assays and molecular beacons, to PCR allows rapid detection of the drug-resistant mutations from smear-positive samples or from culture samples.
PCR also can be used to detect Mycobacterium leprae DNA present in small amounts or to identify the AFB when conventional techniques fail. In addition, PCR is able to detect M leprae after multidrug therapy has been started; because PCR allows early detection of the mycobacteria, treatment can be started even when histopathologic features of the disease are not present [22,43].
Although not FDA approved, PCR can identify different species of Rick-ettsia, such as R rickettsii, Ehrlichia sp, Bartonella henselae (bacillary angio-matosis), and spirochete in late secondary and tertiary cutaneous syphilitic lesions. Among viral pathogens that can be detected by PCR in addition to the ones already mentioned are HSV-1 and -2 from saliva and serum of patients with acute herpes labialis and from spinal fluid; PCR permits detection of either HSV serotype and provides a rapid and definitive diagnosis, has an extraordinary sensitivity greater than virus isolation, or direct detection of HSV antigens or nucleic acids. PCR assays also have established that HSV (mainly HSV-2) is the principal cause of benign recurrent lymphocytic meningitis and confirmed a strong association between HSV-1 infection and Bell's palsy. The clinical significance of small quantities of HSV DNA detected by PCR in the absence of infectious virus remains to be determined. Varicella zoster virus can be identified by multiplex RT-PCR from saliva, tear fluid, skin eruptions, and throat swabs. PCR detects the virus in very early clinical manifestations of the disease with skin biopsy specimens obtained from vesicular and nonvesicular erythematous regions.
In Kaposi's sarcoma the diagnosis can be difficult and PCR can be helpful, because it is highly sensitive and specific for the detection and quantification of human herpes virus 8. PCR is also useful in the detection of human papilloma viruses, such as human papilloma virus-5 (squamous cell carcinomas in renal transplant patients), human papilloma virus-16 (anogenital and cervical malignancies and verrucous carcinoma of the foot), and human papilloma virus-11 (anogenital verrucous carcinoma). PCR also detects other viruses, including parvovirus B19, and also confirms the presence of Epstein-Barr virus in T and B cells, and NK cell lymphomas (accurate classification of the lymphoma). In sudden acute respiratory syndrome, the virus genome was sequenced within weeks of discovery, providing sequence data for the development of RT-PCR based detection strategies.
In relationship to parasitic infections PCR can be used to aid in the diagnosis of Leishmania infantum (cutaneous leishmaniasis). For fungal infections probes also have simplified the identification of the dimorphic fungi, Histoplasma capsulatum, Blastomyces dermatitis, Cryptococcus neoformans, Sporothrix schenckii, Trichophyton rubrum, Coccidioides immitis, and Candida albicans .
The synovial fluid culture is positive in 90% of cases of nongonococcal bacterial arthritis, but Gram stains are positive in only 50% of cases, and clumps of stained or cellular debris may be mistaken for bacteria. Most infected joint effusions are purulent or very inflammatory with average leukocyte counts of 50 to 150,000 c/mL, predominately polymorphonuclear cells. Blood cultures are positive in 50% to 70% of patients with nongonococcal bacterial arthritis. In contrast, the synovial fluid Gram stain is positive in less than 25% of patients with gonococcal arthritis, and culture is positive in only 50%. The skin lesions and blood samples rarely yield positive cultures in disseminated gonococcus infection (DGI). A presumptive diagnosis of DGI is often made on the basis of characteristic signs, symptoms, and identification of N gonorrhoeae from a genitourinary source. Genitourinary cultures are positive in 70% to 90% of patients with DGI. The failure to recover N gonorrhoeae from a site of dissemination may be partly explained by the fastidious in vitro growth requirements of N gonorrhoeae. Immune mechanisms may also be responsible for the sterile synovitis and dermatitis. PCR has been used to detect N gonorrhoeae in patients with clinically typical but culture-negative gonococcal arthritis. The presence of gonococcal DNA, even in culture-negative synovial fluid, suggests that viable bacteria do indeed provoke the synovitis associated with DGI . In a case of staphylo-coccal arthritis, PCR demonstrated persistent Staphylococcus aureus DNA in the synovial fluid for 10 weeks despite adequate antibiotic treatment and sterile synovial fluid [46,47]. Identification of bacterial DNA by PCR is most useful in patients with partly treated or culture-negative bacterial arthritis and in reactive arthritis . PCR in synovial fluid also plays a role in potentially infected joint prostheses.
A healthy but genetically predisposed individual may develop reactive arthritis after a suitable triggering infection. Most commonly the initial infection is localized in the digestive or urogenital tract, even by saprophyte flora; however, the list of microbes able to trigger reactive arthritis is not completely understood and the primary infection may also affect other organs. Despite intensive research, the pathogenic process is not completely understood; however, during the period of contracting the initial infection, incubation time, the primary illness, and the following interval period before the onset, critical immune reactions are thought to take place. A large variety of different microbes lead to a similar clinical entity; isolation of the causative microbe is only rarely possible, and the symptoms and signs of the primary or triggering infection actually may have been quite mild or even passed unnoticed [49,50]. Reactive arthritis is discussed in more detail elsewhere in this issue.
Arthritogenic microorganisms can be detected in synovial samples , not only of patients with reactive arthritis, but also can be found in other autoimmune conditions, such as RA (C trachomatis, C pneumoniae, Mpneumoniae, and Mfermentans) [57-59] and dermatomyositis (parvovirus B19) . With the use of PCR, in the last few years several studies have shown that sometimes more than one microorganism can be present in the same joint [61-63]. This association has been observed for C trachomatis and C pneumoniae, C trachomatis and Borrelia burgdorferi , and for different species of pseudomonas in synovial samples from patients with spon-dyloarthropathy or unexplained arthritis.
Bacterial DNA of C trachomatis and C pneumoniae [65-68]; Yersinia en-terocolitica, Shigella flexneri, and Shigella sonnei; Salmonella typhimurium and Salmonella enteritidis; Pseudomonas sp, Bacillus serius, Campylobacter jejuni, U urealyticum, B burgdoferi, and T whippelii; and RNA of C trachomatis, C pneumoniae, and Yersinia pseudotuberculosis has been found in patients with reactive arthritis and undifferentiated arthritis [69-71]. The search for C trachomatis in the urogenital tract in the first portion of the morning urine by PCR or ligand chain reaction is an acceptable and relatively easy diagnostic approach with a result comparable with urogenital swab analysis. RT-PCR can be used for highly unstable rRNA transcripts to identify viable bacteria in the synovial fluid and synovial tissue samples. Small amounts of sample are needed to perform these tests. PCR is not very sensitive for peripheral blood. There is no agreement on the best technique to detect Chlamydia by PCR, and standardization is still pending. None of the commercially available tests is sensitive enough to detect chlamydial DNA reliably in synovial fluid when compared with the amplification methods, such as nested PCR.
All known bacterial pathogens of humans belong to the eubacteria kingdom. This fact implies that the method to amplify 16S rDNA, such as PCR, could confirm the presence or absence of bacterial pathogens in a normally sterile body. Using this technique, bacterial DNA products seem to be derived from several bacterial species; however, all of them can be found in the human intestinal, urogenital, and respiratory tracts.
In acute anterior uveitis anti-Klebsiella, anti-Y enterocolitica, and anti-Salmonella antibodies have been found [1-5]. In posterior uveitis PCR is increasingly used in the detection of pathogenic organisms associated with many forms of ocular inflammatory and infectious diseases. PCR is able to diagnose viral uveitis, infectious endophthalmitis [72,73], and parasitic eye diseases . The most common identifiable causes of posterior uveitis are infectious agents; in immunocompetent patients Toxoplasma gondii is the most common microorganism identified, whereas in immunosuppressed patients cytomegalovirus, varicella zoster virus, and HSV [74,75] are implicated in acute retinitis . Although local antibody production and viral cultures are useful for the diagnosis, PCR can directly detect the RNA or DNA of the causative microorganism with higher specificity and sensitivity than the other methods [77,78].
Although most of the bacterial causes of conjunctivitis and keratitis are readily cultured or detected by Gram staining, C trachomatis, adenoviruses , herpes simplex, and T whippelii cannot be detected by these methods . PCR primer sets for these microorganisms, however, are able to diagnose these infections in a timely manner. In addition to the detection of infectious diseases of the eye, PCR can also be useful in the diagnosis of B-cell lymphoma that mimics a posterior uveitis and presents as an ocular inflammation in older patients. PCR has demonstrated usefulness for the diagnosis of viral retinitis, conjunctivitis, delayed-onset endophthalmitis, and posterior uveitis [80-82].
NAA assays for the detection of viruses, such as HSV [83-85], cytomegalovirus, enterovirus, hepatitis C and B virus [16,86,87], and HIV [88,89], have proved to be useful for screening, diagnosis, and management. Most PCR assays for viral pathogens have sensitivity of 10 to 100 genomes, which corresponds to less than 1 plaque-forming unit in viral culture . The Canadian Blood Services has adopted NAA methods to screen donated blood for hepatitis C and HIV because of the enhanced sensitivities of these assays. PCR detection of HSV in cerebrospinal fluid has become the gold standard for the diagnosis of herpes encephalitis or meningitis, with sensitivity and specificity of 95% and 94%, respectively, obviating the need for a brain biopsy. Enteroviruses are among the most common causes of aseptic meningitis. PCR for the diagnosis of enteroviral meningitis using cerebrospinal fluid samples has been found to be significantly more sensitive than conventional viral isolation (14% of specimens positive versus 10% positive, respectively) [91-93]. Moreover, the PCR assay can be completed within 1 day, whereas cultures for enteroviruses typically require up to 5 days for isolation of the virus. A PCR assay for cytomegalovirus is available for detection of the virus in plasma or cerebrospinal fluid specimens and has been useful in monitoring HIV and bone marrow transplant patients with cyto-megalovirus infection.
The performance of this test has been comparable with that of antigen assays, with reported sensitivities and specificities of 95% to 98% and
98% to 100%, respectively. In contrast, the sensitivity of culture detection of cytomegalovirus is only 42%. Hepatitis G virus infection can be identified only through PCR testing, which indicates current infection; however, such testing is not readily available or standardized.
In addition to these diagnostic applications, NAA procedures have also been modified to allow for the quantitative measurement of viral load to monitor response to therapy for patients with HIV, cytomegalovirus, or hepatitis C virus infection . For example, measuring HIV viral load in serum has had a major impact on the management of HIV-infected patients. Viral load measurement is of prognostic importance; it is used to predict progression of the disease and to assist in making treatment decisions.
PCR technology has also been used to identify infection by organisms that cannot be cultured. To accomplish this, investigators took advantage of the observation that portions of bacterial 16S rRNA sequences are highly conserved, whereas other regions are less well conserved and are species specific [95,96]. PCR amplification of 16S rRNA sequences of bacteria that cannot be cultured from tissues of patients with such diseases as Whipple's disease and bacillary angiomatosis enabled the discovery and identification of the etiologic agents . Furthermore, using NAA methods, diseases previously thought to be noninfectious have been linked to infectious agents.
The following case scenarios may assist in understanding the clinical usefulness of PCR.
A 19-year-old student is admitted to a local hospital with a 2-week history of fever and monoarthritis involving the knee. Before her admission she had received two courses of oral antibiotic therapy for a presumed upper respiratory infection. Blood and synovial fluid cultures are negative. Should the patient continue with antibiotics? Should antibiotics be stopped because of lack of response?
A 60-year-old man with a 10-year history of RA, who has been receiving tumor necrosis factor antagonists for 2 years, is admitted to the hospital with low-grade fever, malaise, and cough. Three days ago he consulted his primary physician and a purified protein derivative was done that is 10 mm. A chest radiograph indicates the presence of disease in the upper left lobe airspace. Microscopic examination of a sputum specimen reveals a moderate number of AFB. Does this represent tuberculosis or the presence of nontuberculous bacteria?
Each of these clinical scenarios presents the physician with a problem that involves establishing a diagnosis of infection or reactive arthritis in a setting where routine laboratory investigations are likely to be nondiagnostic or do not provide results in a timely manner; however, the PCR will lead them in the right direction.
Was this article helpful?