Increasing incidence of penicillin- and ampicillin-resistant middle ear pathogens.

During a 13-month period ending in January, 1995, we obtained 159 samples of middle ear exudate through tympanocentesis (n = 155) or acute spontaneous otorrhea (n = 4) from 151 children enrolled in therapeutic trials of acute otitis media in a pediatric practice in Northern Virginia. Their ages ranged from < 1 to > 6 years of age (mean, 35 months; median, 22 months). Precise diagnostic criteria for acute otitis media always included bulging outward of all or part of the eardrum, opacification of the eardrum regardless of color and impaired mobility to positive and negative pressure via the pneumatic otoscope. Bacterial pathogens were isolated from middle ear fluid in 95% of these children: Streptococcus pneumoniae was recovered from 61 (37%); Haemophilus influenzae from 45 (27%); Moraxella catarrhalis from 41 (25%); Group A streptococcus from 6 (4%); Staphylococcus aureus from 4 (2%); and no growth or microbes of uncertain significance from 8 (5%). Six of the patients had mixed bacterial cultures; 2 of the 6 had at least one ampicillin-resistant bacteria, and a third had 2 ampicillin-resistant bacteria. Eight patients who failed to improve with antimicrobial treatment had a second tympanocentesis performed or developed spontaneous drainage; on that follow-up culture 3 of 8 cultures had different microorganisms; and 5 of the 8 bacterial specimens were resistant to ampicillin or penicillin. Twenty-one percent of the S. pneumoniae strains recovered from the middle ear were resistant to penicillin. Sixty-two percent of the H. influenzae and 98% of the M. catarrhalis isolates were resistant to ampicillin. Overall bacteria resistant to penicillin or ampicillin were recovered in 54% of middle ear fluid from 46 patients who had received a beta-lactam antibiotic in the preceding month as well as in 57% of middle ear fluids from 105 patients who had not. The empiric use of amoxicillin for treatment of acute otitis media should be reexamined in our community particularly in those who appear ill, have a high fever or have severe unremitting otalgia.

Comparison of an immunochromatographic method for rapid identification of group A streptococcal antigen with culture method.

OBJECTIVES: To compare the sensitivity and specificity of Concise Strep A (Hybritech, San Diego, Calif), an immunochromatographic group A streptococcal rapid antigen detection system, with a two-plate culture method for the diagnosis of streptococcal pharyngitis, and to evaluate the need for routine back-up culture when this rapid test is used. DESIGN: Throat cultures were obtained from 351 children with acute pharyngitis by duplicate rayon-tipped swabs held in parallel and vigorously rubbed against both tonsils and the posterior pharyngeal wall. One swab was tested for group A streptococcal antigen by a registered licensed laboratory technologist in the pediatrician's office. The other swab was streaked over each of two sheep blood agar plates, one of which was enhanced with trimethoprim in combination with sulfamethoxazole. The plain sheep blood agar plate was then incubated in a candle-extinguish jar. The enhanced agar plate was placed in a gas-pack anaerobic jar. Both plates were incubated for up to 48 hours at 35 degrees C. SETTING: A six-person group pediatric practice. PARTICIPANTS: Three hundred fifty-one children. RESULTS: The Concise Strep A antigen detection test produced 129 positive results. Only six of the 129 were not confirmed by culture method. There were four false-negative rapid streptococcal antigen detection test results, all of which were found after a single overnight incubation. The sensitivity for the Concise Strep A test was 96.9% and the specificity was 97.4%. The plain 5% sheep blood agar plate (without trimethoprim and sulfamethoxazole), which was incubated in a candle-extinguish jar, identified 123 (97%) of the 127 positive throat cultures. The second 24-hour incubation and use of trimethoprim and sulfamethoxazole agar were not rewarding for this study. CONCLUSIONS: Concise Strep A, a polyclonal antibody test, in conjunction with a color immunochromatographic assay for soluble streptococcal carbohydrate antigen A appears to be accurate, sensitive, and specific when throat swabs are carefully obtained and when qualified, licensed laboratory technologists perform the procedure. Further studies should be done to confirm our findings, especially when nurses or office staff perform the rapid test procedure in the office setting. If our findings are confirmed, the use of back-up cultures for negative rapid test results obtained using Concise Strep A would be unnecessary.

Evaluation of the EZ-SCREEN enzyme immunoassay test for detection of cocaine and marijuana metabolites in urine specimens.

The EZ-SCREEN Cannabinoid/Cocaine test is a qualitative enzyme immunoassay card test for the detection of cannabinoid and cocaine metabolites in urine specimens. Results are available in about three minutes. The test kit costs approximately $12.00 per two-drug test, and there are no additional expenses for analytical equipment. We compared the results of the EZ-SCREEN test with results of specimens analyzed by gas chromatography/mass spectrometry (GC/MS) for 36 clinical specimens positive for cannabinoids (marijuana), 38 clinical specimens positive for benzoylecgonine (cocaine), and 33 drug-free specimens. Results from our study for marijuana included: sensitivity = 92%; specificity = 89%; and efficiency = 95, 67, and 87%, respectively. In the present study, there were five false-positive tests for cocaine and one false-positive test for cannabinoids. As with all screening test results, confirmation of positive specimens by a more specific quantitative analytical method such as GC/MS is recommended.

Federal guidelines for marijuana screening should have lower cutoff levels. A comparison of results from immunoassays and gas chromatography-mass spectrometry.

Federal guidelines for urine marijuana testing use a screening cutoff of 100 ng/mL and a confirming cutoff of 15 ng/mL. We tested 75 urine samples with two different immunoassays with cutoff points of 100 and 20 ng/mL. The same samples were also analyzed with gas chromatography-mass spectrometry. At the government recommended cutoff levels for screening and confirming, test sensitivity was 47% and specificity was 91%. With a screening cutoff of 20 ng/mL confirmed at 5 ng/mL, test sensitivity was 88% and specificity was 94%. We suggest adoption of these lower cutoff levels.

Evaluation of colorimetric dipstick test to detect alcohol in saliva: a pilot study.

The Alco Screen Saliva Dipstick is an inexpensive, easy-to-use, colorimetric test that gives a semiquantitative estimation of the blood alcohol value by measuring the relative concentration of salivary alcohol. To evaluate its accuracy, we compared the results from tests with the Alco dipstick with values from simultaneously measured blood alcohol tests in 53 patients who were suspected of having ingested alcohol. The correlation between Alco dipstick results and blood alcohol values was strong (r [Spearman's rho], + .91). When the blood alcohol concentration was 0.1 g/dL or more, the Alco dipstick test was 90.9% sensitive, 71.4% specific, and 92% efficient. At Alco dipstick values of 0.02 and 0.05 g/dL, however, semiquantitative concordance was unsatisfactory. Nevertheless, even at the 0.05-g/dL value of salivary alcohol, the test was still valuable as a screen of de facto alcohol ingestion. Definitive diagnosis of relative alcohol intoxication requires confirmatory breath or blood alcohol concentrations by standard methodologies.

Evaluation of the Keystone Diagnostic Quik Test. A paper chromatography test for drugs of abuse in urine.

The Keystone Diagnostic Inc (KDI) Quik Test is advertised to be an on-site, rapid, broad-spectrum paper chromatography test that is marketed for use in industry, drug abuse treatment facilities, and physician's offices. Even when unskilled individuals perform the test, five drugs of abuse, excluding marijuana, are said by the manufacturer to be reliably detected. In this study, ten confirmed drug-free urine specimens and 56 specimens positive by an immunoassay screening test, and confirmed by gas chromatography/mass spectrometry, for cocaine (n = 32), phencyclidine (n = 12), or opiate drugs (n = 12), were reanalyzed by the KDI Quik Test. The sensitivity of the KDI Quik Test was 62.5%, specificity was 70%, predictive value of a negative test was 36%, and efficiency was 64%. In too many instances the KDI Quik Test did not detect the drugs for which it is intended. Paper chromatography tests such as the KDI Quik Test are inaccurate and as such are unacceptable in any setting for the purposes of screening for drugs of abuse.

Urinary cannabinoids in monitoring abstinence in a drug abuse treatment program.

Urine specimens from 162 adolescents entering a drug treatment program were tested for cannabinoids using a photometric immunoassay (EMITst) and thin-layer chromatography (TLC). The EMITst has a cutoff point of detection 25 ng/mL or less of 9-carboxy-tetrahydrocannabinol. When reported as positive, both tests appeared to be reliable. There were two false-positive EMITst results and three false-positive TLC results in the 67 urine specimens that did not contain cannabinoids. When reported as negative, however, the EMITst with its 100 ng/mL cutoff failed to detect almost 40% of all cannabinoid-positive specimens. Of the 65 chronic marijuana smokers in the present study who stated that they had smoked within two days of their admission into the treatment facility, 17 (26%) went undetected by the 100 ng/mL cutoff used by the EMITst method. In clinical settings such as drug treatment programs, tests for urinary cannabinoids should use a detection threshold at 20 ng/mL or less.