Methods of investigation in clinical cardiology. X. Studies on the evaluation of diagnostic tests in cardiology

Diagnosis is a complex cognitive process which is characterised by uncertainty. This uncertainty can be managed through specific knowledge in conjunction with probability theory. Studies evaluating diagnostic tests are the best way of building this knowledge. Studies evaluating diagnostic tests have two essential components: the gold standard and the new test. Both components, gold standard and test, are independent measurement process that can be influenced by diverse sources of variability. The comparison between diagnostic and test is essentially a hierarchical procedure. Diagnostic tests are evaluated by their sensitivity, specificity compared to a definitive gold standard. The predictive values and the likelihood ratio test are also used. Sensitivity (the proportion of true positives) and specificity (the proportion of true negatives) are values obtained from a sample and thereby can be considered as the conditional agreement between gold standard and the new test. Kappa coefficients for sensitivity and specificity are useful tools for adjusting both indices. Sensitivity and specificity are non-poblational values, they are estimates of the true values of the study population and can be affected by random error and systematic errors (bias). Confidence intervals are useful for giving an indication of the precision of the point estimates of sensitivity and specificity. A suitable sound design is required to avoiding a biased estimate of sensitivity, likelihood ratio, and predictive values. Finally a list of potential biases is given with methods for minimising these.     

Helicobacter eradication: an expensive Sisyphus task

BACKGROUND: The benefits of eradicating H. pylori infection vary about the antimicrobial regimen. In contrast, comparing many clinical studies the overall outcomes seem independent of the kind of treatment. Therapy depends on accurate diagnosis. Therefore, the observed differences and correspondences can be caused by not obviously but systematic influences using noninvasive diagnostic tests. METHOD: Simulations concerning noninvasive tests with variable specificity and sensitivity were performed. The results were compared to the hypothetical estimate that 1/3 of the population of Germany were H. pylori infected. This group should be traced and checked again after a treatment with 90% efficacy. RESULTS: In comparison to the a priori conditions most decisions according to noninvasive tests were not valid. Both sensitivity and specificity differ significantly from 100%, therefore the accuracy was low. The differences between true positives and true negatives were unpredictably accelerated if the inclusion criteria depend on the poor sensitivity of the first test. For the second test which is commonly used regarding its specificity, there is a change in the statistical behavior. As a result, the homogeneity that is true and false negatives and positives of the non-infected or infected groups differed considerably. One of the examples was a procedure using a test with 79% specificity and 85% sensitivity. If the prevalence of the infection was 33% the simulations demonstrated a therapeutic effect of about 76.5% instead of a priori efficacy of 90%. Under these conditions 14% of the population become false positive. Furthermore, the simulations revealed a second effect. Using the same test with low accuracy twice the investigator got the impression of superior therapeutic considerations. The misleading conclusion was caused by diminished sensitivity if the specificity remained constant. Thus, neither serological tests nor the 13C-breathtests can be assumed as validated methods for both to screen infected people and to check the treatment. Furthermore, the noninvasive tests are not useful for the determination of the prevalence of H. pylori infection. The influence of variable sensitivity and specificity of a diagnostic test on the outcome of a clinical trial can become a bias or can be used to manipulate the outcome. CONCLUSION: The effect of a therapeutic regimen depends on the prevalence of the H. pylori infection. Because it is difficult to separate true positives, in clinical trials the main effects and interactions can be improved under the conditions used in "experimental design". Some earlier results must be reevaluated because of the bias resulting from inadequate diagnostic tests.     

Head-to-head evaluation of five chlamydia tests relative to a quality-assured culture standard

Nucleic acid amplification tests offer superior sensitivity for the detection of Chlamydia trachomatis infection, but many laboratories still use nonamplification methods because of the lower cost and ease of use. In spite of their availability for more than a decade, few studies have directly compared the nonamplification tests. Such comparisons are still needed in addition to studies that directly compare individual nonamplification and amplification tests. The purpose of this study was to evaluate and compare the performance characteristics relative to culture of five different tests for the detection of C. trachomatis with and without confirmation of positive results. The tests were applied to endocervical specimens from 4,980 women attending family planning clinics in the northwestern United States. The five nonculture tests included Chlamydiazyme (Abbott), MicroTrak direct fluorescent antibody (DFA) (Syva), MicroTrak enzyme immunoassay (EIA) (Syva), Pace 2 (Gen-Probe), and Pathfinder EIA (Sanofi/Kallestad). All positive results obtained with a nonculture test (except MicroTrak DFA) were confirmed by testing the original specimens with a blocking antibody test (Chlamydiazyme), a cytospin DFA (MicroTrak EIA and Pathfinder EIA), and a probe competition assay (Pace 2). The prevalence of culture-proven chlamydia was 3.9%. The sensitivities of the nonculture tests were in a range from 62 to 75%, and significant differences between tests in terms of sensitivity were observed. The positive predictive value for each test was 0.85 or higher. The specificities of the nonculture tests without performance of confirmations were greater than 99%. Performing confirmatory tests eliminated nearly all of the false positives.     

Increased anticoagulation during cardiopulmonary bypass by prostaglandin E1

We prospectively evaluated the diagnostic accuracy of an epinephrine-containing epidural test dose (EpiTD) as a marker of intravascular injection in 209 unmedicated laboring women. Maternal heart rate (MHR) was continuously monitored and recorded on a strip chart. A tocodynamometer monitored uterine activity. A lumbar epidural catheter was placed and aspirated. If aspiration was positive for blood or cerebrospinal fluid (CSF), the catheter was replaced. In uterine diastole and with stable MHR, 198 patients received an EpiTD (epinephrine 15 microg plus lidocaine 45 mg) via the catheter. MHR and the generated HR strip were observed. A positive EpiTD was defined as a sudden increase in MHR of 10 bpm more than the resting MHR, within one minute after the injection, with a fast acceleratory phase of more than 1 bpm. Absence of a tachycardiac response suggested a negative EpiTD. If the tachycardiac response was deemed equivocal or a uterine contraction followed the EpiTD injection within 1 min, the EpiTD was invalidated and repeated. Catheter aspiration was repeated, and the catheter was removed if aspiration was positive. All patients with negative EpiTD and aspiration received 6-12 mL of epidural bupivacaine 0.25% with or without fentanyl 50 microg. Absence of analgesia without signs or symptoms of systemic toxicity after a maximum of bupivacaine 30 mg defined failed epidural analgesia. All patients with positive EpiTD and negative aspiration received 5 mL of lidocaine 2% epidurally as a second test dose (Lido100TD). The presence of tinnitus and/or metallic taste defined a positive Lido100TD. There were 176 true negatives, 0 false negatives, 14 true positives, and 8 false positives. The sensitivity of EpiTD was 100%, the specificity 96%, the negative predictive value 100%, and the positive predictive value 63%. The prevalence of negative tests was 88%, and the prevalence of positive tests was 12%. The overall accuracy of an EpiTD was 95.5%. We conclude that EpiTD is a reliable test to identify i.v. catheters during the performance of lumbar epidural analgesia in laboring patients. Implications: Catheters inserted for epidural analgesia in laboring patients may accidentally enter a blood vessel. Local anesthetics injected through these catheters may cause seizures and cardiac arrest. In this study, we concluded that injecting a small amount of epinephrine before injecting a local anesthetic frequently helps to identify these misplaced catheters. Few catheters may actually be in the correct place even after responses to epinephrine.     

Hidden Markov models for detecting remote protein homologies

MOTIVATION: A new hidden Markov model method (SAM-T98) for finding remote homologs of protein sequences is described and evaluated. The method begins with a single target sequence and iteratively builds a hidden Markov model (HMM) from the sequence and homologs found using the HMM for database search. SAM-T98 is also used to construct model libraries automatically from sequences in structural databases. METHODS: We evaluate the SAM-T98 method with four datasets. Three of the test sets are fold-recognition tests, where the correct answers are determined by structural similarity. The fourth uses a curated database. The method is compared against WU-BLASTP and against DOUBLE-BLAST, a two-step method similar to ISS, but using BLAST instead of FASTA. RESULTS: SAM-T98 had the fewest errors in all tests-dramatically so for the fold-recognition tests. At the minimum-error point on the SCOP (Structural Classification of Proteins)-domains test, SAM-T98 got 880 true positives and 68 false positives, DOUBLE-BLAST got 533 true positives with 71 false positives, and WU-BLASTP got 353 true positives with 24 false positives. The method is optimized to recognize superfamilies, and would require parameter adjustment to be used to find family or fold relationships. One key to the performance of the HMM method is a new score-normalization technique that compares the score to the score with a reversed model rather than to a uniform null model. AVAILABILITY: A World Wide Web server, as well as information on obtaining the Sequence Alignment and Modeling (SAM) software suite, can be found at http://www.cse.ucsc.edu/research/compbi o/ CONTACT: karplus@cse.ucsc.edu; http://www.cse.ucsc.edu/karplus     

The clinical assessment of cancer tests: a statistical approach to the prediction of "false positive" rates from controls

As the sensitivity of the new in vitro cancer detection techniques increases, a larger number of "false positive" results can be expected from control populations apparently free from malignant disease at the time of the test. An attempt is made to predict the number of unexplained positives, together with the expected age and sex distribution, using published cancer registration statistics. The predicted numbers are compared with the observed numbers of unexplained positives in two clinical trials of the MEM test.     

The detection of fake-bad and fake-good responding on the Millon Clinical Multiaxial Inventory III.

The purpose of this study was to examine the effectiveness of the 3 Modifying Indices of the Millon Clinical Multiaxial Inventory III (MCMI—III) in the detection of fake-bad and fake-good responding. The sample consisted of 160 psychiatric outpatients. Paired t tests were performed to examine the effects of instructional set (faking vs standard instructions). As hypothesized, instructional set produced significant differences on Scale X, Scale Y, and Scale Z in both fake-bad and fake-good analyses. Single-scale cutoff scores were as effective as multiple-scale cutoffs. The overall rates of successful classification indicated moderate effectiveness and utility of the MCMI—III Modifying Indices in the detection of dissimulated responding. When base rates were varied to more closely approximate a general clinical population, overall classification accuracy increased, but identification of faking (positive predictive power) gradually eroded with declining base-rate estimates. At lower base rates of faking, MCMI—III standard cutoff points yielded a high number of false positives. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Application of anticancer agent embolism in Miles'' operation

A double-blind prospective study was done with 15 patients with anterior shoulder instability to determine the diagnostic efficacy of magnetic resonance (MR) imaging versus arthroscopy in the evaluation of chondral or osteochondral lesions of the humeral head. MR produced 6 true positives, 5 true negatives and 4 false negatives, and its accuracy and sensitivity were 60% and 87%, respectively, whereas arthroscopy gave 8 true positives, 5 true negatives and 2 false negatives, with a sensitivity of 80% and an accuracy of 87%. All lesions diagnosed with either method were regarded as positive by definition, with the result that the specificity was always 100%. The differences in diagnosis sprang from the false negatives. The 40% discrepancy between the two methods was probably due to our distinction in MR between intra- and extra-articular osteochondral lesions. In the first group (the 4 MR false negatives), there were three instances of first-degree intra-articular lesion and one diagnostic error (third-degree lesion). In the second (the 2 arthroscopy false negatives), the lesions were of the extra-articular type. It is thus advisable to employ both of these methods to ensure the correct diagnosis of a Hill-Sachs lesion, and hence the correct choice of treatment.     

Finding intron/exon splice junctions using INFO, INterruption Finder and Organizer

INFO, INterruption Finder and Organizer, has been used to find coding sequence intron-exon splice junctions in human and other DNA by comparing the six conceptual translations of the input DNA sequence with sequences in protein databanks using a similarity matrix and windowing algorithm. Similarities detected both delineate position of the gene and provide clues as to the function of the gene product. In addition to use of a standard similarity matrix and windowing algorithm, INFO uses two novel steps, the MiniLibrary and Reverse Sequence steps, to enhance identification of small exons and to improve precision of junction nucleotide delineation. Exons as small as about 30 bases can be reliably found, and > 90% of junctions are precisely identified when canonical splice junction information is used. With the MiniLibrary and Reverse Sequence steps, INFO parameters need not be optimized by the user. In comparative test runs using 19 human DNA sequences, INFO found 108 of 111 exons, with 0 reported false positives, compared with 111 exons and 51 false positives for BLASTX, 99 exons and 6 false positives for GRAIL II, 77 exons and 24 false positives for GeneMark, 61 exons and 9 false positives for GeneID, and 105 exons and 6 false positives for PROCRUSTES. The correlation coefficient for finding and positioning these 111 exons was greater than 98% for INFO. Comparable results were obtained in test runs of 13 nonhuman DNA sequences. INFO is applicable to DNA from any species, will become more robust as sequence databanks expand, and complements other heuristic approaches.     

Enzyme tests in diseases of the prostate

Enzyme tests in diseases of the prostate focus primarily on the use of serum acid phosphatase assays in patients with suspected or histologically proved adenocarcinoma of the prostate. The purpose of this review is to consider various potential clinical uses of the assay, to examine the data available concerning the performance of the test in given clinical situations and to define those situations in which the test is actually useful. Included in this discussion will be sources of false positive and false negative values, predictive values of the test in clinical settings and efforts to minimize shortcomings.     

Cockroach allergy: a study of its prevalence using skin tests with commercial extracts

Cockroaches have been increasingly recognized as an important source of indoor allergens. In this study we assessed the prevalence of cockroach sensitization among an outpatient population observed at our Department (155 patients with a mean age of 29 +/- 12 years), based on skin "prick" tests with four commercial cockroach extracts along with a common battery of standardized inhalant allergens. We found a positive wheal to at least one of these four extracts in 27 patients: 26 (96.2%) to Blatta orientalis, 10 (37%) both to Blatella germanica and 1 (3.7%) exclusively to Blatella germanica, with no significant concordance between them. We also observed in patients with cockroach positive skin "prick" test an association with atopy (p < 0.001) and with cutaneous reactivity to other indoor allergens, namely house dust mites (p = 0.02), danders (p = 0.01) and fungi (p = 0.01). These data confirm the higher risk of cockroach sensitization among the atopic population sensitized to indoor allergens. However, the heterogeneity of the positive cutaneous responses obtained in this study, possibly reflecting the incomplete standardization of cockroach extracts, questions the real prevalence and clinical significance of this particular sensitization.     

Occurrence of introns in the 16S rRNA genes of members of the genus Thermoproteus

Cervical smears (n = 150) from five departments showing high-grade dyskaryosis were examined by three cytologists. All the smears came from patients with biopsy-proven CIN III. One hundred had been correctly reported (true positives) but 50 had originally been reported as negative and had been found to be positive only on review (false negatives). There were significant differences between the two sets in the characteristics of the dyskaryotic cell population. The false-negative smears tended to have fewer than 200 dyskaryotic cells. The nuclei of the dyskaryotic cells tended to have fine rather than coarse nuclear chromatin. A smear with fewer than 50 dyskaryotic cells is 26 times more likely to be reported as negative than one with more than 200 dyskaryotic cells. The results suggest that there is a type of severely dyskaryotic smear that is inherently likely to be missed on routine screening.     

Experiences with the tuberculin tine test in serial tests

During the past year the tuberculin tine test, frequently used in Switzerland to indicate tuberculin allergy, was again the subject of controversy. A retrospective study analysed results of tuberculin tests in school children (average age 15) obtained by the Basle School Health Service during 1976--1978. A total of 4919 tine tests were read during this period. 2416 of the total of 3405 tine-negative pupils were also tested with 10 TU PPD Berna according to Mantoux. 357 (14.8%) showed positive reactions. Since not all tine-negative pupils were submitted to a Mantoux test, it must be assumed that 14.8% false negatives were missed in the group not further tested. On this basis, a sensitivity of 74.4% was calculated for the tine test. It is a debatable point whether this is sufficient for practical use.     

Diagnosis of Trypanosoma evansi in Saudi Arabian camels (Camelus dromedarius) by the passive haemagglutination test and Ag-ELISA

The passive haemagglutination test and Ag-ELISA were employed to monitor antibody titres and antigenaemia levels in 4 Najdi camels experimentally infected with Trypanosoma evansi. The two tests were also used to determine the prevalence of trypanosomiasis in a total of 218 Najdi camels in the Gassim region, Central Saudi Arabia, during the period from October 1992 to September 1993. Trypanosoma evansi antibodies in the experimentally infected camels rose after 14-21 days and reached a maximum of between 1:64 and 1:128 by the 12th week post infection. Circulating antigens were detected in the experimentally infected camels one week post infection and antigenaemia levels fluctuated but generally remained above preinfection OD values. The results obtained from the field survey showed that 5.5% of the sampled camels were parasitologically positive for trypanosomes, while 19.7% were serologically positive by the passive haemagglutination test and 13.8% by Ag-ELISA. No significant age difference in seropositivity was observed in the tested camels.     

Evaluation of 2-SP transport medium for detection of Chlamydia trachomatis and Neisseria gonorrhoeae by two automated amplification systems and culture for chlamydia

AIMS: To assess the performance of 2-sucrose-phosphate based transport medium (2-SP) for the detection of Chlamydia trachomatis and Neisseria gonorrhoeae by an automated commercial polymerase chain reaction (PCR) and ligase chain reaction (LCR) compared to centrifugation culture on McCoy cells for C trachomatis. Second, to compare both amplification systems for initial diagnostic testing of a low prevalence population for sexually transmitted diseases. METHODS: Four hundred and eighty one consecutive urogenital and conjunctival specimens were examined. All tests were performed on the same specimen collected with a dacron swab and transported in 2-SP medium. Samples that were positive by culture or by both PCR and LCR were considered to be true positives. RESULTS: The prevalences of C trachomatis and of N gonorrhoeae were 2.7% and 0.4%, respectively. PCR had a resolved sensitivity and specificity of 100% and 99.8%, respectively, for C trachomatis, and 100% and 98.9%, respectively, for N gonorrhoeae. LCR was 100% sensitive and specific for both pathogens. The resolved sensitivity of the shell vial assay was 85%. No culture positive sample would have been missed by PCR or LCR. The inhibition rate for PCR was 4.8%. CONCLUSIONS: 2-SP medium proved to be suitable for both PCR and LCR. It is not limited to any one test manufacturer and allows the performance of amplification techniques and viral and chlamydia culture from the same specimen. The LCR was more reliable than PCR on initial testing. However, hands on time is longer, and no amplification control is available for LCR.     

Evaluation of an enzyme immunosorbent assay for the diagnosis of Argentine haemorrhagic fever

To elaborate a set of serological tests for the diagnosis of Argentine haemorrhagic fever (AHF), an enzyme-linked immunosorbent assay (ELISA) for detection of specific anti-Junin virus (JV) IgG is described, and its performance is compared with that of the plaque reduction neutralization test (PRNT). The reproducibility, sensitivity, specificity, and confidence limits for positive and negative results for ELISA were statistically analysed. The value of 800 was demonstrated as the lowest positive titer. Titers > or = 800 varied within one (two-fold) dilution in 95.6% of the tests, while the sensitivity and specificity were 99.2% and 98.8%, respectively. The assay yielded 1% of false positives and 0.05% of false negatives. A comparison of ELISA to PRNT in detecting the seroconversion for JV was studied by the chi square test (comparison of proportions in paired samples) and the K parameter for agreement proportion. Comparison of ELISA to PRNT showed no significant difference in the proportions of positive and negative results of these assays (P < 0.01), demonstrating an equivalent performance (K = 0.98) in the diagnosis of AHF. In addition, the simplicity and safety of the procedures involved make this ELISA the most suitable test to detect natural human JV infections.     

The sensitivity and specificity of the ergometric test for the diagnosis of coronary disease

According to the principles of probabilistic analysis, sensitivity and specificity of a diagnostic test are fixed values. Nevertheless, most authors consider them to be inconstant values, specially when applied to the diagnosis of coronary heart disease by exercise stress test. In this paper, we review the basic concepts on sensitivity and specificity of diagnostic tests and try to explain their supposed variability, when related to exercise test, as a function of undue comparison between ST-segment response and the findings of cinecoronariography. Based on the essential difference between coronary heart disease and ischemic heart disease, we demonstrate why such an equivocal comparison can lead to false results of sensitivity and specificity of exercise tests relative to coronary heart disease. As a result, their alleged variability depends most on the prevalence of ischemia throughout the spectrum of coronary heart disease in the studied population. As a matter of fact, unless one can rely on a method as a gold standard for the diagnosis of ischemic heart disease, the real sensitivity and specificity of exercise stress test should be considered as unknown values.     

Novel mutations and DNA-based screening in non-Jewish carriers of Tay-Sachs disease

We have evaluated the feasibility of using PCR-based mutation screening for non-Jewish enzyme-defined carriers identified through Tay-Sachs disease-prevention programs. Although Tay-Sachs mutations are rare in the general population, non-Jewish individuals may be screened as spouses of Jewish carriers or as relatives of probands. In order to define a panel of alleles that might account for the majority of mutations in non-Jewish carriers, we investigated 26 independent alleles from 20 obligate carriers and 3 affected individuals. Eighteen alleles were represented by 12 previously identified mutations, 7 that were newly identified, and 1 that remains unidentified. We then investigated 46 enzyme-defined carrier alleles: 19 were pseudodeficiency alleles, and five mutations accounted for 15 other alleles. An eighth new mutation was detected among enzyme-defined carriers. Eleven alleles remain unidentified, despite the testing for 23 alleles. Some may represent false positives for the enzyme test. Our results indicate that predominant mutations, other than the two pseudodeficiency alleles (739C-->T and 745C-->T) and one disease allele (IVS9+1G-->A), do not occur in the general population. This suggests that it is not possible to define a collection of mutations that could identify an overwhelming majority of the alleles in non-Jews who may require Tay-Sachs carrier screening. We conclude that determination of carrier status by DNA analysis alone is inefficient because of the large proportion of rare alleles. Notwithstanding the possibility of false positives inherent to enzyme screening, this method remains an essential component of carrier screening in non-Jews. DNA screening can be best used as an adjunct to enzyme testing to exclude known HEXA pseudodeficiency alleles, the IVS9+1G-->A disease allele, and other mutations relevant to the subject's genetic heritage.     

Ergometric tests in cardiology. Diagnosis of coronary insufficiency: methods and evaluated parameters

The historical development of the stress-test in the diagnosis of coronary insufficiency has been examined from the first observations about 1930 regarding changes in the repolarization phase during effort in coronary patients, up to modern tests with the ergometer bicycle and treadmill. Starting from the consideration that Master's Test is still the most commonly used in clinical practice, the limitations of tests of this type are highlighted and the discussion also covers the techniques and parameters now considered of greatest importance in cardiopathy diagnosis and evaluated by means of modern maximal stress tests. The results of a first period of work involving tests using the treadmill are reported. The methodology is discussed and the symptoms or ECG data that had suggested the test be used are related to the patient's origin (out-patient or hospitalized) and with the test's positivity or negativity. The high incidence of unstable ST syndrome, especially in the female sex, is also stressed. If this is not thoroughly investigated functionally (hyperventilation, Valsalva, etc.) it could be the cause of a large number of false positives. The lack of danger in the maximal stress test, even in cardiopaths, is confirmed together with the extreme ease with which nearly all patients manage to perform the test on the treadmill. Stress is also laid on the fact that the stress test is functional, unlike coronarography which is purely morphological, and the two examinations are thus complementary in the diversity of information they provide.     

Double immunochemical screening test (hemoglobin and albumin) for the detection of occult intestinal bleeding

BACKGROUND AND OBJECTIVES: Chlamydia trachomatis antigen testing of clinical specimens is replacing culture as the test of choice. Because of a potential for false positive results in low prevalence populations, there is an apparent need for confirming specimens positive by enzyme immunoassay (EIA). GOAL OF THIS STUDY: To examine specimens falsely positive in the Chlamydiazyme EIA assay according to gender and specimen type. STUDY DESIGN: Testing of genitourinary specimens from men and women consecutively enrolled from five health care delivery sources in an urban Canadian population. All specimens were initially tested in the Chlamydiazyme test and all positives repeated in a confirmatory blocking assay provided by the manufacturer. Additional confirmatory testing was performed using immunofluorescence (IF) staining for C. trachomatis elementary bodies (EB's) and polymerase chain reaction (PCR). RESULTS: From Jan. 1, 1990 to June 1, 1991, multiple specimens from 656 men and 5,628 women of varying population prevalences were screened. EIA-positive specimens from women had a repeat negative rate of 22% to 27% from cervical swabs and 29% from urethral swabs. Male urethral swabs had a high repeat negative rate of 22% when EIA was the only positive test, but 2.4% when the specimen was positive by EIA and culture. EIA-positive first void urine (FVU) specimens from men had a repeat negative rate of 8.7% as opposed to 17% to 32% from women. Only 1.7% (2/115) of male FVU did not block compared to rates of 47% (22/47) to 80% (4/5) in FVU from women. Analysis of EIA optical densities (OD's) and EB counts showed an association between the absorbance range 0.1 to 1.4 OD and 0-85 EB's. The greatest number of EB's and highest OD's were seen with cervical specimens, followed by urine and urethral specimens in women infected at all three specimens. All 55 specimens that did not confirm in the blocking test had no EB's and a convenience sample of seven were negative by PCR. All of a subset of 50 blocked specimens contained EB's or were positive by PCR. CONCLUSIONS: Although a variable proportion of specimens may not repeat positive in the EIA, use of the blocking reagent to confirm the repeat positives is highly recommended and the rate of blocking may be heavily influenced by gender and specimen type.