Rapid Detection Methods inMicrobiology: Are They Right for Your Office?

Office Practice of Laboratory Medicine

0025-7125/87 $0.00

+ .20

Rapid Detection Methods in Microbiology: Are They Right for Your Office? Cynthia A. Needham, Ph.D. *

Infection represents a common problem in outpatient medicine, and the clinician is generally confronted with a patient whose expectations for prompt diagnosis and treatment do not match the realities of available test information. Traditionally, a large percentage of these infections involve the upper respiratory, urinary, or genital tract, where culture has represented the gold standard by which therapeutic decisions are made. Unfortunately, culture results are not available for several days after the patient ha,s been evaluated-well beyond the time frame when decisions are made. As a result, antibiotic choice is either delayed or empiric. Within the last several years, the use of such technical advances as monoclonal antibodies has allowed the development of a new generation of tests for diagnosing infectious diseases. The use of these tests has made it possible to document rapidly the presence of specific bacterial, fungal, and viral pathogens in clinical material. Some of these tests are now available in a format that can be applied to office practice. Many of these tests still require a relatively high level of sophistication on the part of the individual performing them. However, the benefits that can derive from the timely provision of diagnostic information probably outweigh the difficulties of quality assurance for the tests. Like all diagnostic tests, the new generations of rapid methodologies have inherent levels of reliability. It is extremely important that the characteristics of these tests be understood within the context of the disease setting. Because it would be difficult to list exhaustively all of the new methods available, the focus of this article will be to give examples of tests *Associate Professor, Boston University School of Medicine, Boston; and Section Head, Microbiology, Department of Laboratory Medicine, Lahey Clinic Medical Center, Buriington, Massachusetts This article was adapted from Needham CA: Rapid methods in microbiology for in-office testing. Clin Lab Med 6(2):291-304, 1986.

Medical Clinics of North America-Vo!' 71, No. 4, July 1987

591

592

CYNTIIIA

A.

NEEDHAl'.1

and discuss the impact of false-positive and -negative results within each clinical setting. INFECTIONS OF THE UPPER RESPIRATORY TRACT Infections of the upper respiratory tract are the most common infectious diseases seen in the primary care setting. It is estimated that they result in over 15 million office visits per year in the United States and cost $300 million to treat. 18 The most common etiologic agents are the respiratory viruses, which cause a mild, self-limited disease whose symptoms frequently include pharyngitis. 34 Currently, the approach to these infections is merely supportive, because there are no effective antiviral agents available for treating them. This is in contrast to bacterial pathogens for which effective antibiotics do exist. Unfortunately, there is significant overlap in symptoms between these two groups, and it becomes a major therapeutic challenge to the physician to distinguish between viral and bacterial agents of pharyngitis. The most common cause of bacterial pharyngitis is the group A streptococcus (GAS), causing 15 to 30 per cent of cases. Children 5 to 10 years of age are the population at risk, with peak incidence occurring during the first few years of school. Although the role of specific antibiotic therapy in alleviating symptoms of GAS pharyngitis remains controversial, its role in the prevention of noninfectious complications of the disease (acute rheumatic fever) does not. For years, a throat culture has served as the only test for the diagnosis of GAS pharyngitis. However, there are some serious limitations to this traditional approach, particularly when attempting to move the methodology into the office setting. The minimum time necessary to recover and identify the organism is 24 hours. More typically, culture results are not available for 48 to 72 hours after the patient is seen. A relatively high level of expertise is required to perform the culture in a reliable manner. Several reports have suggested that accuracy achieved in an office setting may be as low as 40 to 50 per cent when compared with results obtained from a reference or hospital-based laboratory. 22, 26 In addition, physiologic variation among strains of GAS complicates their detection by routine methods. The most notable example of this is the existence of strains that are not betahemolytic when grown in O 2 and consequently not detected unless cultures are incubated anaerobically. 28. 29 These strains may represent up to 5 per cent of total isolates. The use of direct detection methods may well eliminate all three of these limitations. Direct detection methods rely on the recognition of a specific cell wall-associated antigen of the GAS by an antibody directed against it. The antigen is exposed by treating the cells with either an enzymatic or chemical digestion. Either of these procedures can be performed directly on the swab used to collect the specimen, thus obviating the need for growing the organism. Once the antigen has been released, it can be reacted with the specific antibody. Depending on the digestion method and the detection method for the antigen-antibody complex, results

593

RAPID DETECTION METHODS IN MICROBIOLOGY

Table 1. Tests for Direct Detection of Group A Streptococci SENSITIVITY METHOD (MANUFACTURER)

ELISA Ventrascreen Strep A (Ventre x Laboratories, Inc.; Portland, ME) Latex agglutination Culturette Brand 10-Minute Group A Strep ID (Marion Laboratories, Inc.; Kansas City, MO) Directigen Group A Strep Test (Hynson, Wescott, and Dunning; Baltimore, MD) Group A Strep Direct Antigen Identification Test (Difco Laboratories; Detroit, MI) Detect-A-Strep (Antibodies Inc.; Davis, CA) Streptex (Wellcome Diagnostics; Research Triangle Park, NC) PathoDx Strep A (Diagnostic Products Corp.; Los Angeles, CA)

SPECIFICITY

(REFERENCES)

(%)

TOTAL TEST TIME (MIN.)

Chemical

90*

91 *

20

Chemical

90-95 (9. 19, 39, 57)

93-100

10-12

Enzymatic

72-91 (36,38)

96-99

65

Enzymatic

83 (62)

99

60

Chemical

87 (5)

99

10

Enzymatic

86

99

70

Chemical

79

98

6-10

EXTRACTION PROCEDURE

(%)

*Manufacturer's data.

are available from 10 minutes to 1 hour after the specimen has been collected, The methods are easy to perform and do not require extensive training to interpret. With the use of appropriate control materials, the methods are reliable and have the additional advantage of detecting strains of GAS that are not beta-hemolytic. Direct detection kits that are commercially available and for which evaluations have been published are listed in Table 1. There are a plethora of new assays that will soon be available in addition to those listed. These include products from Murex (SUDS Group A Strep Test; Atlanta, GA), Hygeia (Flow-Through Immunosorbent Assay; Newton, MA), Hybritech (Strep A ICON; San Diego, CA), and Abbott (Test-Pak A; North Chicago, IL). Early reports indicate that the new products will perform as well as those now available, and several of the methods have positive and negative quality control built into the assay. The tests differ in the method of cellwall digestion and the linkage chemistry and carrier material used for the antibody. Enzymatic digestion, in general, takes longer than chemical digestion. If the antibody is linked to an enzyme, it is an enzyme-linked immunosorbent assay (ELISA) test, and the end-product is determined via a colorimetric reaction. If the antibody is linked to a latex particle, it is a latex agglutination test, and the end-product is macroscopic clumping of

594 the latex particles. If the antibody is link~d to a fluorescein mole~ule, it is a fluorescent assay, and the end-product IS read as a fluorescent SIgnal. When one examines the performance characteristics of each of these assays, it is clear that some patients will have a positive culture and a negative direct test for GAS (false negative), and some patients will have a negative culture and a positive direct test (false positive). GAS pharyngitis is a treatable disease with potential morbidity and mortality associated with inappropriate therapy. At the same time, the consequence of a false-positive test is that a patient without GAS would be treated with antibiotics that are generally low in cost and have limited untoward effects. Given that most of the direct tests have a specificity of 90 per cent or greater, their use should reduce the number of patients given empiric therapy. The rate for anaphylaxis following penicillin therapy is no more than 0.025 per cent. With a false-positive rate of 5 per cent, there would be less than one case of anaphylaxis for 80,000 patients treated on the basis of a direct antigen test. Similarly, it is estimated that the asymptomatic carriage rate of GAS may be as high as 10 per cent, which might be equatible to a false-positive culture. The consequence of a false-negative result is a more difficult issue. In general, it appears that the higher the microbiologic burden, the more likely one is to get a positive result by direct testing. It is tempting to assume that the presence of small numbers of GAS recovered in culture probably represent colonization rather than active disease. Unfortunately, this has not been borne out when evidence of serologic conversion has been sought in patients with positive cultures. In a recent study, Gerber and Randolf2° demonstrated that 36 per cent of patients with low numbers of GAS and falsely negative direct tests for the organism still had a significant antibody rise. CoinCidentally, they also found that only 44 per cent of patients with both tests positive experienced a similar increase in antibody titer, further substantiating that a positive test and/or large numbers of organisms recovered does not differentiate colonization from disease. The failure to treat GAS pharyngitis adequately puts the patient at risk for developing acute rheumatic fever. However, the incidence of acute rheumatic fever has decreased dramatically over the last decades. The incidence of acute rheumatic fever in physicians' practices serving relatively affiuent, middle-class Americans approximates 0.5/100,000. Thus, given the small number of patients who hav.e a false-negative test and serologic evidence of disease, the sensitivity of direct testing appears to be acceptable. This is particularly true when one examines the unreliability of traditional culture techniques when performed by unskilled individuals. In addition, it is always possible to perform a standard culture on those patients with negative direct tests and signs and symptoms consistent with GAS pharyngitis.

INFECTIONS OF THE URINARY TRACT Urinary tract infections are the second most common bacterial infection seen in ambulatory medicine. It has been estimated that between 10 and

RAPID DETEC:TIO~ 'vlETIIODS IK 'vlICROBIOLOGY

595

20 per cent of women will expcrience a urinary tract infection at some time during their lives."o Most of these infections are uncomplicated, easy to treat, and pose no risk for the patient in terms of morbidity or mortality. Exceptions to this observation are infections occurring in children, men, pregnant women, and those persons experiencing recurrcnt or relapsing infection. The diagnosis of urinary tract infection still relies on the demonstration of bacteria in significant numbers in the urine. However, substantial controversy exists as to the number of bacteria needed to qualify as "significant." Early studies of Kass·33 and Sanford and associates 5 ! established the level of significance as greater than 100,000 colony-forming units (CFU) per milliliter of urine. This decision level was derived from data collected by culturing the first-morning, clean-voided, midstrcam urine specimens from asymptomatic women. Recent evidence suggests that as many as one half of women presenting with dysuria and coliform bacteria in the bladder may have counts as low as 100 CFU per ml in both voided and bladder urine. 60 In addition, the relative state of hydration and the time since last voiding will greatly influence the number of organisms detected, making lower numbers of bacteria potentially significant. Unfortunately, lowering the decision level for significant bacteriuria will also increase the number of specimens with small numbers of contaminating bacteria present that do not constitute significant infection. The physician who is performing his or her own laboratory testing is in an enviable position, because the decision to screen for significant bacteriuria using some of the newer methods can be weighed against the patient's clinical setting. In general, the threshold for significant bacteriuria in asymptomatic, ambulatory patients is greater than 100,000 CFU per ml, but the threshold for significant bacteriuria in symptomatic patients may be in the range of 100 to 100,000 CFU per m!. As will be pointed out in this section, various screening methods become increasingly unreliable when fewer bacteria are present, and thus have limited use in patients in whom lower counts may be significant. Rapid Methods for Detection of Urinary Tract Infection A rapid test to detect significant bacteriuria would be very useful because symptoms are not always present, and when present they may result from diseases other than urinary tract infection. Unfortunately, almost all of the rapid tests currently available have a very high rate of false positivity and, consequently, a very poor positive predictive value. In general, they are not useful to identify patients with bona fide infections. On the other hand, there are several that are quite sensitive, even when lower numbers of bacteria are considered as significant, and the prcdictive value of a negative is quite good. This means that they may bc useful in identifying patients who do not have urinary tract infections and who do not require a culture. The rapid screens for significant bacteriuria differ from those for a specific organism such as the group A streptococcus. They are primarily based on some measure of biomass that is not specific for any organism. The least expensive and least labor intensive of the rapid tests is the

596

CYNTHlA

A.

NEEDHA~I

Table 2. Rapid Screening Methods for Detecting Significant Bacteriuria and/or Pyuria*

PREDICTIVE VALUE

SENSITIVITY ~IETHOD (~IA:-.iUFACTURER)

(%)

Negatice (%)

66--88

42-54

91-95

0.17

59-78

43-61

96-97

0.60

67-95

33-76

91-98

1.20

68-97

39-92

92-98

0.10

SPECIFICITY

(REFERENCES)

74-78 Chemstrip LN (Bio-Dynamics; (27,46, 47) Indianapolis, IN) 93-94 Bac-T-Screen (Marion (3, 15, 45, 48) Laboratories, Inc.; Kansas City, MO) 88-94 Lumac Biocounter (3M Co.; (15, 27, 37, 64) St. Paul, MN) 84-94 Gram stain (11,48)

LIST COST

Positice (%)

(%)

PER TEST

($)

*Results presented in the table are the compilation of reported studies and are based on a decision level of greater than 10' colony-forming units per milliliter for significant bacteriuria.

Chemstrip LN (Boehringer-Ylannheim Diagnostics, Bio-Dynamics; Indianapolis, IN). This is a plastic strip to which reagent papers for indicating the presence of leukocyte esterase and nitrite in urine are attached. Granulocytic leukocytes contain esterases that catalyze the hydrolysis of a substrate that is impregnated in the reagent in the paper to produce a purple color, thus indicating the presence of leukocytes in the specimen. The nitrite test depends on the conversion of nitrate (derived from the diet) to nitrite by the action of bacteria in the urine. If nitrite is present in the urine, it reacts with the reagents in the test strip to produce a pink color. The test takes approximately 2 minutes to perform and costs approximately $0.17 per test. Not all patients with pyuria have urinary tract infections, and not all bacteria that cause urinary tract infections can metabolize nitrate and produce nitrite. Nonetheless, the use of the two tests together results in a negative predictive value of greater than 90 per cent when the level of significant bacteriuria is set at greater than 100,000 CFU per ml (Table 2). Of course, the number of false negatives increases substantially when a lower colony count is considered significant. Thus, its primary use may be in screening asymptomatic patient populations to eliminate performing cultures on those persons with negative tests. It should be emphasized that the leukocyte esterase test is a sensitive method of detecting significant pyuria (greater than 10 leukocytes per mm 2) and as such can be used reliably to eliminate unnecessary urine microscopy in the majority of patients,'54. 61 It should also be noted that if the two tests are both positive, the LN strip has a much better positive predictive value. Although a culture still probably needs to be performed, the detection of white blood cells and nitrite is sufficient information to initiate therapy empirically. The Bac-T-Screen (Marion Laboratories, Inc.; Kansas City, MO) represents a slightly different approach to urine screening. This method employs a relatively inexpensive instrument that draws a milliliter of urine through a special filter. Bacteria and leukocytes in the specimen will adhere

RAPID DETECTION METHODS IN ,'vI!CHOBIOLOGY

597

to the filter, which is subsequently stained and decolorized. The filter is then placed in an attached card reader that measures the amount of residual stain against a background constant. A positive result can be expected for both significant bacteriuria and pyuria, but the test cannot distinguish one from the other. The Bac-T-Screen can be performed in approximately 2 minutes. The cost per test is approximately $0.60 and the filter apparatus with attached card reader costs $1700. The major advantage of the Bac-T-Screen is its ability to detect lower levels of bacteriuria more reliably than most other urine screens, particularly in association with pyuria. For example, in a multicenter clinical trial conducted at three separate sites,12 the Bac-T -Screen was able to detect 86.5 per cent of specimens containing greater than 100 CFU per ml and greater than 10 white blood cells per mm 3 . The disadvantage to this level of sensitivity is the concomitant loss of specificity, with the "false-positive" rate reportedly as high as 30 to 40 per cent. 45. 46 Thus the Bac-T -Screen can be used to screen both asymptomatic and symptomatic patients but should not be used in either group to determine therapy because of the high falsepositive rate. Screening by bioluminescence is a third alternative. There are several instruments that can perform this type of screen, but the 3M Lumac Biocounter (3M Co.; St. Paul, MN) is probably the most cost effective in terms of reagents and technical time. The technology is based on the principle that both bacterial and nonbacterial cells from the host are present in the urine and contain ATP. This ATP can be detected by the bioluminescence produced in the firefly luciferin-Iuciferase reaction: ATP + luciferin luciferas~ adenyl luciferin + 0 2 adenyl oxyluciferin + light. Early work with this approach was discouraging because of interference from nonbacterial sources of ATP. However, the Lumac kit contains reagents that selectively isolate and destroy the mammalian ATP before the actual assay of bacterial ATP. Clinical evaluations of the Lumac have demonstrated that the test is comparable to the two previous methods (see Table 2). As a screen for negative urine specimens, the Lumac has a negative predictive value well over 90 per cent, even when lower colony counts are considered significant. However, two distinct disadvantages to this test method are the turnaround time and the cost of the instrument and reagents. Although the hands-on time for the assay is only 4 minutes, the complete assay involves several steps and requires about 25 minutes to generate a single result. Obviously, this method could be made efficient by batching specimens, but then it also loses its appeal as a rapid test method. The list price of the instrument is approximately $8000, and the average cost of reagents is $1.20 per test. These two considerations (time and expense) argue against using this method in the physician's office setting. Although the purpose of this article is to highlight new developments in office testing, the Gram-stained smear must be included in any discussion of rapid screens for significant bacteriuria. The smear is performed by placing one drop of uncentrifuged urine on a slide. Using a criterion for positivity of greater than three bacteria per oil immersion field (1000 X

CYNTHIA A. NEEDHAM

598

g 'fication) the sensitivity, specificity, and predictive value of positive

m~ n~gative :esults compare quite favorably with the values of other rapid :et~ods. Furthermore, the smear may yield valuable information as to the

specimen quality as well as the most probable pathogen, and it still represents the least costly test i~ ~erms of reagents. . Because the positive predIctIve value of most of these tests IS not particularly good, it is di~cult. to justif~ .their use .in the decision to treat a presumed urinary tract mfectIon empIrIcally. WIth some of the screens, approximately one half of the patients with a positive result would not have infection and would thus be treated unnecessarily, possibly obscuring other causes for their physical symptoms. Negative predictive value tends to be better for all of the screens. However, one must weigh carefully the consequences of an untreated urinary tract infection in any given patient population. When prevalence is high, there will be a relatively high absolute number of patients with urinary tract infections who will not be identified by rapid screening methods.

SEXUALLY TRANSMITTED DISEASES The third large group of infections encountered in outpatient medicine are sexually transmitted diseases. Although the list of etiologic agents has expanded rapidly over the last decade, three specific pathogens account for the largest percentage of the total infections: Neisseria gonorrhoeae, Chlamydia trachomatis, and Herpes simplex virus (HSV). Each of these agents represents a slightly different challenge to the laboratory in terms of diagnostic testing. Each will be discussed in terms of rapid methods that might be appropriate in the physician office practice. Rapid Detection of N. Gonorrhoeae Gonorrhea still ranks first among the reportable diseases in the United States. There are usually about 1 million cases per year reported to the Centers for Disease Control, and it is estimated that an equal number may go unreported. 8 Persons 20 to 24 years of age account for 35 to 40 per cent of reported cases each year, and persons 15 to 19 years of age for nearly 25 per cent. 7 Even though morbidity has declined over the last several years, gonorrhea represents a major public health problem because of its incidence. The most common site of localized infection in both male and female patients is the genital tract. In the male patient, infection usually presents as urethritis and in the female patient as cervicitis. In general, infection in the male patient is usually symptomatic and results in few complications. In contrast, infection in the female patient is more frequently asymptomatic and has a higher rate of complications. N. gonorrhoeae is a difficult organism to recover in culture. The organism is highly susceptible to environmental conditions and will not survive if stringent transport and/or culture requirements are not met. Isolation and identification may require several days to complete. The gonococcus is consequently a prime candidate for direct test methods not

RAPID DETECTION METHODS IN MICROBIOLOGY

599

dependent on growth. There is currently only one specific direct test method available-an ELISA test called Gonozyme (Abbott Laboratories, North Chicago, IL). Gonozyme is a solid-phase assay that will detect a specific gonococcal antigen in swabs collected from the urethra or cervix. To perform the assay, the antigen is eluted from the swab and adsorbed to a specially treated bead. After the bead is washed to remove unbound material, it is reacted with rabbit antibody specific for the gonococcus, followed by peroxidase-conjugated antirabbit immunoglobulin G. Finally, a peroxidase substrate is added, and a visible color reaction develops if the specimen is positive. The intensity of the color is measured spectrophotometrically. In order to assess the utility of this test, it is easiest to look at specific patient groups in whom Gonozyme has been tested. The assay has the highest performance standards when used to diagnose symptomatic male patients with urethritis. 23, 35, 43, 44, 52, 59 The reported sensitivity of the test in this population is greater than 93 per cent. Very few false positives have been documented, and specificity is greater than 95 per cent. Thus, the direct test method would appear to be a reliable alternative for diagnosing gonorrhea in the symptomatic male patient. It must be remembered, however, that the Gram stain is an accepted method of diagnosing disease in this population. In the hands of a skilled microscopist, the Gram stain has the same sensitivity and specificity as the direct antigen test. 58 The ELISA procedure is relatively labor intensive, costly, and requires about 3 hours to perform. It is difficult to justify its use in the symptomatic male population when the alternative is both technically simple and cost effective. Unfortunately, the Gonozyme test does not fare as well in the remaining patient groups: asymptomatic male and female patients, and symptomatic female patients. The reported sensitivity in those groups is significantly less than with traditional culture techniques, detecting only about 75 per cent of specimens with subsequently positive cultures. The specificity remains high for male patients but becomes alarmingly low in endocervical secretions, with false positive results occurring in as many as 23 per cent of women tested. 23 Given the social implications of a positive test for gonorrhea, this level of specificity is unacceptable, particularly when screening low risk populations. In addition, the test cannot be used for sites other than the urethra and endocervix, necessitating the use of another method for these specimens. Although conceptually a direct test for N. gonorrhoeae is very appealing, this test in its present configuration cannot be recommended for general use. Another test that is available for the presumptive diagnosis of gonococcal urethritis in men is Gonodecten (U.S. Packing Corp.; LaPorte, IN). To perform this test, urethral discharge is collected on a swab, placed in a plastic tube, and moistened with oxidase reagent. Development of a purple color on the swab within 3 minutes is a positive test for the presumptive diagnosis of gonorrhea. In the one published evaluation of this product, Gonodecten had a sensitivity of 95.6 per cent compared with culture, but a specificity of only 84.2 per cent. 30 In the same study, the Gram stain was 99.5 per cent sensitive and 100 per cent specific compared with culture. Because of the low specificity and limited scope of this test, it does not

600

CYNTHIA A. NEEDHA~1

offer any advantage over the Gram-stained smear in the diagnosis of gonorrhea. " One final limitation in the use of dIrect detectlOn methods for N. gonorrhoeae always must be . r~~embered. ~one of .the assays is abl~ to predict the presence of pemcIllmase-producmg strams of the orgamsm. Particularly in areas of the country where antibiotic resistant strains are prevalent, it is preferable to culture for the organism so that its susceptibility to penicillin can be documented. Rapid Detection of C. Trachomatis

C. trachomatis may be the most common agent of sexually transmitted disease in the United States today. The epidemiology and clinical spectrum of the disease parallel that of N. gonorrhoeae. 53 The bacterium is an obligate intracellular parasite and, like viruses, requires cell culture for its recovery in vitro. Needless to say, this is a procedure that is sufficiently sophisticated to be outside the realm of most hospital laboratories. Recently, there have been two' direct methods made available for detection of C. trachomatis that have simplified testing for the organism's presence in clinical specimens. Chlamydiazyme (Abbott Laboratories; North Chicago, IL) is an ELISA procedure that can be performed directly on swab specimens collected from urethral, cervical, rectal, and conjunctival sites. The test procedure is equivalent to that described previously for Gonozyme, with a total test time of 4 hours. The sensitivity of the method seems to be slightly better in specimens collected from female than from male patients. Most studies report a sensitivity of approximately 90 per cent for women and 73 to 80 per cent for men. 2 . 10. 16.32.49 Specificity, on the other hand, is quite high in both male and female populations, exceeding 95 per cent in all reports to date. MicroTrak (Syva Company; Palo Alto, CA) is an alternative to the ELISA test for direct detection of C. trachomatis. To perform the test, slides are prepared using a swab from the urethra or endocervix. The slide is fixed and then stained using a fluorescein-Iabeled monoclonal antibody that is specific for the organism. If C. trachomatis is present, it can be detected by examining the slide with a microscope equipped with a fluorescent light source and optics. The test requires approximately 1 hour of processing time, personnel well trained in fluorescent microscopy, and a microscope with excellent fluorescent optics. The sensitivity and specificity of the MicroTrak direct specimen test compare favorably with that of both cell culture and the Chlamydiazyme test. In most reports, the sensitivity exceeds 90 per cent and the specificity is greater than 95 per cent. 2 . 10. 14. 31 The MicroTrak test may have some advantage over Chlamydiazyme in that specimen quality can be evaluated simultaneously with examination of the smear for the presence of the microorganism. However, this may be offset by the degree of skill required of the person performing the microscopy. In contrast to gonorrhea, there are no simple procedures to document infection with C. trachomatis. Tissue culture is complex and time consuming, with a final result not available until 4 to 14 days after specimen

RAPID DETECTION METHODS IN :'viICROBIOLOGY

601

processing. However, even though the direct tests represent an improvement over conventional methods, the decision to perform them in office practice should be weighed carefully. The tests are relatively complicated technically and are not sufficiently rapid to generate a result while the patient is still in the office. In addition, both MicroTrak and Chlamydiazyme require the purchase of relatively expensive equipment, which for the moment has limited application. Most reference laboratories offer one or both of these direct testing methods, and turnaround time for results is usually 24 hours. Unless a high volume of testing is anticipated, it would seem preferable to refer specimens rather than provide testing on site. It also must be remembered that the predictive value of a positive will be reduced when the incidence of disease is very low in a population. This is true for both of the direct tests for C. trachomatis. Although there is less of a social stigma attached to this disease than to gonorrhea, the use of these tests in a low-prevalence population will lead to more people being treated unnecessarily. Rapid Detection of HSV Infection Genital herpes is an increasingly important sexually transmitted disease. It is estimated that 5 million adults in the United States were infected as of 1980, and that an additional 250,000 will become infected annually. 42 Recent estimates from the Centers for Disease Control suggest that visits to private physicians for genital herpes have increased from 29,500 in 1966 to 260,000 in 1979. 6 Although the epidemiologic features are slightly different from gonorrhea, the age group at greatest risk is similar. Precise and rapid diagnosis of genital herpes infection is desirable to permit the use of newly available antiviral agents, counseling of patients in terms of infectivity and long-term risk for cervical cancer, and the prudent management of pregnant women. The gold standard against which all other methods must be compared is isolation of the virus in cell culture. Although HSV is one of the easiest and most rapid viruses to isolate, tissue culture may not provide timely results. The examination of Giemsa-stained smears (Tzanck preparations) of scrapings from the base of vesicles or ulcers for the presence of typical viral cytopathic effect has served traditionally as an adjunct to culture. The sensitivity of the Tzanck preparation ranges from as low as 50 per cent in recurrent disease to as high as 80 per cent in primary lesions. 17, 41 Apart from relatively low sensitivity, there are some other difficulties associated with this procedure, Artifact can be easily introduced if the smears are not prepared carefully and fixed immediately. The cytopathologic changes are not specific, necessarily, for the HSV, and the microscopist must have a reasonable level of skill and experience in reading the smears. Direct immunofluorescent staining of material taken from active HSV lesions can be performed using monoclonal antibodies, and this offers a specific, relatively rapid approqch to diagnosis. When active disease is present, the diagnostic yield has been reported to be as high as 85 to 88 per cent. However, in the absence of active clinical disease, this method becomes less sensitive and the reading of smears increasingly difficult. An alternative direct test methodology, an ELISA test (Ortho Diagnostic;

CYNTHlA A. NEEDHAM

602

Carpinteria, CA), is also available. The assay is performed in microtiter wells that have been precoated with anti-HSV antibody, which traps antigen eluted from the specimen swab. The trapped antigen is then reacted with a peroxidase-conjugated antibody to HSV followed by peroxidase substrate. The entire procedure requires approximately 5 hours of processing time and involves several manipulations during that time interval. Unfortunately, the sensitivity of the ELrSA has been disappointingly low when compared with tissue culture. Reported sensitivity has been as low as 45 per cent,25 but more typically is between 60 and 70 per cent. 40, 56, 63 In the same studies, specificity has been excellent, with most investigators reporting results approaching 100 per cent. Thus, whereas the specificity is improved over that of the Tzanck preparation, the test is much less sensitive, and at this time much more labor intensive. Additionally, there are substantial equipment requirements. All of these factors currently preclude the use of this method in the physician's office laboratory. A latex agglutination test (Virogen Herpes Slide Test, Wampole Laboratories; Cranbury, NY) has recently become available. The test employs a disposable card slide onto which is placed a 50-I-LI sample that has been eluted from the swab used to collect the specimen. Latex particles coated with antiherpes antibody are added and the card placed on a rotator for 20 minutes. A positive reaction is read as agglutination of the latex particles. The test requires approximately 25 to 30 minutes to perform. Early evaluations of the test have not been promising, with investigators reporting a sensitivity between 21 and 49 per cent compared with tissue culture. Specificity, on the other hand, is quite high, and is comparable with either the fluorescent or ELISA method. The manufacturer recommends that the test be used only if wet lesions are present for sampling, and thus far the test has not been evaluated in a setting where specific clinical information was available. Because this test is substantially easier than either of the previously described direct test methods, it still remains a promising alternative for the office setting. CONCLUSION There are now a variety of rapid test methods available to assist in the diagnosis of the three most common infectious diseases seen in ambulatory medicine: pharyngitis, urinary tract 'infection, and venereal disease. Before choosing to implement any of these methods, several aspects should be considered carefully. The objectives met by on-site testing should be identified clearly and the appropriate test chosen to meet those objectives. For example, if the goal is to establish a diagnosis while the patient is still available in the office, a test method that requires more than 20 to 30 minutes to perform will not fulfill the objective. Similarly, if tests must be batched for maximum efficiency, timeliness of results will be compromised. The prevalence of the specific agent or disease in the patient population should be established, as prevalence will significantly influence the predictive value of a diagnostic test. The lower the prevalence of a disease, the

RAPID DETECTION METHODS IN MICROBIOLOCY

603

lower the predictive value of a positive result and the higher the predictive value of a negative result. Conversely, the higher the prevalence, the higher the positive predictive value and the lower the negative predictive value. Thus, in one setting a test may serve best as a screen to rule out disease and, in another setting, as a confirmatory test to establish a diagnosis. The impact of falsely positive and/or falsely negative results must be clearly appreciated to avoid diagnostic errors. Finally, the technical features of a test method should be appraised carefully. How difficult is the test to perform, and what skill level is necessary to ensure the validity of results? How much quality control is necessary? How much additional equipment is needed to perform the assay, and does the equipment have multiple applications or just a single application? Does the cost-benefit ratio of providing on-site testing exceed that of referring the specimens to a reference laboratory?

REFERENCES 1. Barr RJ: Rapid method for Tzanck preparations. JAMA 237:1119-1120, 1977 2. Baselski V, McNeeley G, Robinson M: Comparison of two noncultural tests to culture for detection of Chlamydia trachomatis in obstetric patients (abstr. no. 727). Intersci Conf Antimicrob Agents Chemother 1985 3. Bixler-Forell E, Bertram MA, Bruckner DA: Clinical evaluation of three rapid methods for the detection of significant bacteriuria. J Clin Microbiol 22:62-67, 1985 4. Brown ST: Sensitivity and specificity of diagnostic tests for genital infection with Herpesvirus hominis. Sex Transm Dis 6:10-16, 1979 5. Carlson JR, Huston AC, Watson TE, et al: Development and evaluation of Detect-AStrep: A new rapid latex agglutination kit for the direct detection of Group A streptococcal antigen in throat swabs (abstr. no. 15). Intersci Conf Antimicrob Agents Chemother 1984 6. Centers for Disease Control: Genital herpes infection-United States, 1966-1979. MMWR 31:137, 1982 7. Centers for Disease Control: Gonorrhea-United States, 1983. MMWR 33:361-363, 1983 8. Centers for Disease Control: STD Fact Sheet. Edition 35. U.S. Dept. Health and Human Services pub!. no. 81-8195. Atlanta, Centers for Disease Control, 1981 9. Chang MJ, Mohla C: Ten-minute detection of group A streptococci in pediatric throat swabs. J Clin Microbiol 21:258-259, 1985 10. Chernesky M, Leman C, Mahoney J, et al: A comparison of cell culture, Chlamydiazyme, and MicroTrak performed on genital specimens from high and medium risk populations (abstr. no. 726). Intersci Conf Antimicrobiol Agents Chemother 1985 11. Crout F, Tilton RC: Rapid screening of urine for significant bacteriuria by Gram stain, acridine orange stain, and the Autobac MTS system. Diagn Microbiol Infect Dis 2:179-186, 1984 12. Daifuku R, Zeleznik D, Stamm WE: Detection of bacteriuria and pyuria with the Bac-TScreen bacteriuria detection device (abstr. no. 1072). Intersci Conf Antimicrobiol Agents Chemother 1984 13. Daly JA, Rufener ML, Gooch WM, et al: Outpatient evaluation of Group A streptococcal pharyngitis in children using a rapid, direct test from throat swabs (abstr. no. 195). Intersci Conf Antimicrob Agents Chemother 1985 14. Deddish PA, Khurana CM: Evaluation of the MicroTrak direct specimen test for the detection of Chlamydia trachomatis in cervical specimens (abstr. no. 725). Intersci Conf Antimicrob Agents Chemother 1985 15. Drow DL, Baum CH, Hirschfield G: Comparison of the Lumac and Monolight systems for detection of bacteriuria by bioluminescence. J Clin MicrobioI20:797-801, 1984 16. England B, Howard L, Laux R, et al: Clinical evaluation of Chlamydiazyme for the

604 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

CYNTHIA A. NEEDHA'"

detection of chlamydial urogenital infection (abstr. no. 723). Intersci Conf Antimicrobiol Agents Chemother 1985 Fife K, Corey L: Laboratory diagnosis of sexually transmitted diseases: Viral agents. In Holmes KK, Mardh P, Sparling PF, Wiesner PJ (eds): Sexually Transmitted Diseases. New York, McGraw-Hill, 1984 Fischer PM, Addison LA, Curtis P, et al: The Office Laboratory, Norwalk, AppletonCentury-Crofts, 1983 Gangemi JD, Gibson J, Krech L, et al: Clinical evaluation of a latex agglutination test for Group A streptococcus in throat swab specimens (abstr. no. C59). Annu Meeting Am Soc Microbiol 1985 Gerber MA, Randolf MF: Sensitivity of antigen detection tests for Group A beta-hemolytic streptococcal pharyngitis with respect to bona fide infections (abstr. no. 194). Intersci Conf Antimicrobiol Agents Chemother 1985 Ginsberg CM, McCracken GH, Steinberg JB, et al: Treatment of Group A streptococcal pharyngitis in children. Clin Pediatr 21:83-88, 1982 Glass R, Fischer Z, Bottone EJ: Inaccuracy of house staff in reading throat cultures. JAMA 240:2651-2652, 1978 Granato PA, Roefaro M: Comparative evaluation of enzyme immunoassay and culture for the laboratory diagnosis of gonorrhea. Am J Clin Pathol 83:613-618, 1985 Gulick P, .Hall G, McHenry MC: Office microbiology. Med Clin North Am 67(1):39-56, 1983 Hoffman B, Jungkind D, Hailer G, et al: Evaluation of two rapid methods for the detection of Herpes simplex virus antigen in patient specimens (abstr. no. C238). Annu Meeting Am Soc Microbiol 1985 Hoffman S: Lack of reliability of primary grouping of beta-hemolytic streptococci by culture of throat swabs with Streptocult supplemented with bacitracin discs in general practice. J Clin Microbiol 22:497-500, 1985 Hughes JG, Synder RJ, Washington JA: An evaluation of a leukocyte esterase/nitrite test strip and a bioluminescence assay for detection of bacteriuria. Diagn Microbiol Infect Dis 3:139-142, 1985 Hull KH, Wolsh LJ: Use of three screening methods for detecting streptococcal pharyngitis (abstr. no. C336). Annu Meeting Am Soc Microbiol 1985 James L, McFarland RB: An epidemic of pharyngitis due to a nonhemolytic Group A streptococcus at Lowry Air Force Base. N Engl J Med 284:750-752, 1971 Janda WM, Jackson T: Evaluation of Gonodecten for the presumptive diagnosis of gononococcal urethritis in men. J Clin Microbiol 21:143-145, 1985 Jaqua-Stewart JJ, Tichota-Lee J, Jaqua RJ: Chlamydia trachomatis-rapid testing by direct immunofluorescence (abstr. no. 721). Intersci Conf Antimicrobiol Agents Chemother 1985 Jones MF, Smith TF, Houglum AJ, et al: Detection of Chlamydia trachomatis in genital specimens by the Chlamydiazyme test. J Clin Microbiol 20:465-467, 1984 Kass EH: Asymptomatic infections of the urinary tract. Trans Assoc Am Physicians 69:56-61, 1956. Levy ML, Ericsson ED, Pickering LK: Infections of the upper respiratory tract. Med Clin North Am 67(1):153-172, 1983 Manis RD, Harris B, Geiseler PJ: Evaluation of Gonozyme, an enzyme immunoassay for the rapid diagnosis of gonorrhea. J Clin MicrobioI20:742-746, 1984 Marsik F, Owens D: Comparison of the Directigen method to culture for the detection of Streptococcus pyogenes from throat swabs (abstr. no. C58). Annu Meeting Am Soc Microbiol 1985 Martin ET, Cote JA, Perry LK, et al: Clinical evaluation of the Lumac bioluminescence method for screening urine specimens. J Clin Microbiol 22:19-22, 1985 McCusker JJ, McCoy EL, Young CL, et al: Comparison of Directigen Group A Strep Test with a traditional culture technique for detection of Group A beta-hemolytic streptococci. J Clin Microbiol 20:824-825, 1984 Miceika BG, Vitom AS, Thompson KD: Detection of Group A streptococcal antigen directly from throat swabs with a ten-minute latex agglutination test. J Clin Microbiol 21:467-469, 1985 Morgan M, Smith TF: Evaluation of an enzyme-linked immunosorbent assay for the detection of Herpes simplex virus antigen. J Clin Microbiol 19:730-732, 1984

RAPID DETECTIOl\: :\1ETIIODS IN :\1ICROBIOLOGY

605

41. Motyl MR, Botton~ El, Janda JM: Diagnosis of Herpesvirus infections: Correlation of Tzanck preparations with viral isolation. Diagn Microbiol Infect Dis 2:157-160, 1984 42. Nahmias AI, Josey WE: Herpes simplex viruses 1 and 2. In Evans AS (ed): Viral Infections of Humans. Edition 2. New York, Plenum, 1982 43. Papasian Cl, Bartholomew WR, Amsterdam D: Validity of an enzyme immunoassay for detection of Neisseria gonorrhoeae antigens. J Clin Microbiol 19:347-350, 1984 44. Papasian CJ, Bartholomew WR, Amsterdam D: Modified enzyme immunoassay for detecting Neisseria gonorrhoeae antigens. J Clin Microbiol 20:641-643, 1984 45. Pezzlo MT, Wetkowski MA, Peterson EM, et al: Detection of bacteriuria and pyuria within two minutes. J Clin Microbiol 21:578-581, 1985 46. Pfaller MA, Koontz FP: Use of rapid screening tests in processing urine specimens by conventional culture and the AutoMicrobic System. J Clin Microbiol 21:783-787, 1985 47. Pfaller MA, Koontz FP: Laboratory evaluation of leukocyte esterase and nitrite tests for the detection of bacteriuria. J Clin Microbiol 21:840-842, 1985 48. Roe C, Needham CA: Diagnostic value and cost-effectiveness of a new screening method in the detection of bacteriuria (abstr. no. 187). Intersci Conf Antimicrob Agents Chemother 1985 49. Ryan R, Steingrimsson 0, Kwasnik I, et al: The detection of chlamydial antigen by Chlamydiazyme in both high risk and lower risk populations (abstr. no. 724). Intersci Conf Antimicrobiol Agents Chemother 1985 50. Sanford JP: Urinary tract symptoms and infections. Annu Rev Med 26:485, 1975 51. Sanford JP, Favour CB, Mao FH: Evaluation of the "positive" urine culture. Am J Med 19:88-91, 1956 52. Schachter I, McCormack WM, Smith RF, et al: Enzyme immunoassay for diagnosis of gonorrhea. J Clin Microbiol 19:57-59, 1984 53. Schachter I, Crossman M: Chlamydial infections. Annu Rev Med 32:45-61, 1981 54. Schumann CB, Creenberg NF: Usefulness of macroscopic urinalysis as a screening procedure. Am J Clin Pathol 71:452-456, 1979 55. Schwartz RH, Wientzen RL, Pedreira F, et al: Penicillin V for Croup A streptococcal pharyngotonsillitis: A randomized trial of seven vs. ten days' therapy. JAMA 246:1790-1794,1981 56. Sewell DL, Horn SA: Evaluation of a commercial enzyme-linked immunosorbent assay for the detection of Herpes simplex virus. J Clin Microbiol 21:457-458, 1985 57. Slifkin M, Cil CM: Evaluation of the culturette brand ten-minute group A strep ID technique. J Clin Microbiol 20:12-14, 1984 58. Sparling PF: Neisseria gonorrhoeae. In Mandell CL, Douglas RC, Bennett JE (eds): Infectious Diseases and Their Etiologic Agents. New York, John Wiley & Sons, 1979 59. Stamm WE, Cole B, Fennell C, et al: Antigen detection for the diagnosis of gonorrhea. J Clin Microbiol 19:399-403, 1984 60. Stamm WE, Counts CW, Running KR, et al: Diagnosis of coliform infection in acutely dysuric women. N Engl J Med 307:463-468, 1982 61. Statland BE: Macroscopic analysis of urine. Med Lab Observ 15:13-15, 1983 62. Venezia R, Ryan A, Alward S, et al: Evaluation of a rapid method for the detection of streptococcal Croup A antigen directly from throat swabs. J Clin Microbiol 21:395-398, 1985 63. Warford AL, Levy RA, Rekrut KA: Evaluation of a commercial enzyme-linked immunosorbent assay for detection of Herpes simplex virus antigen. J Clin MicrobioI20:490-493, 1984 64. Welch WD, Thompson L, Layman M, et al: Evaluation of two bioluminescence-measuring instruments, the Turner design and Lumac systems, for the rapid screening of urine specimens. J Clin Microbiol 20:1165-1170, 1984 Department of Laboratory Medicine University Hospital Boston University Medical Center 75 E. Newton Street Boston, Massachusetts 02118