Population-based prospective surveillance of invasive pneumococcal disease was done in Southern California from 31 March 1992 to 1 April 1995; 814 cases were identified, for an incidence of 12.5/100,000 persons/year. The incidence among persons < or = 2, < or = 5, and > or = 65 years of age was 145, 72, and 32/100,000, respectively. More than 95% of cases included bacteremia; incidence of meningitis was 0.8/100,000. Among children < or = 2 years of age, 79% of isolates were obtained in the outpatient setting, compared with 16% of isolates among persons > or = 15 years of age. Eighty percent of isolates were serotypes included in heptavalent pneumococcal conjugate vaccines currently being evaluated. Children < or = 2 years of age were at highest risk of having an isolate resistant to penicillin. Among resistant isolates, high-level resistance increased from 4% to 21% over a 3-year period. Prospective epidemiologic data are needed to perform a protective efficacy trail of pneumococcal conjugate vaccines in infants, among whom most invasive pneumococcal disease is vaccine-preventable.

We evaluated the BACTEC MGIT 960 system, which is a fully automated, noninvasive system for the growth and detection of mycobacteria with a capacity to incubate and continuously monitor 960 7-ml culture tubes. We studied 3,330 specimens, 2,210 respiratory and 1,120 nonrespiratory specimens, collected from 2,346 patients treated at six sites. Processed specimens were inoculated into the BACTEC MGIT 960 and BACTEC 460 TB systems, as well as onto Lowenstein-Jensen slants and Middlebrook 7H11/7H11 selective plates. From all culture systems, a total of 362 isolates of mycobacteria were recovered; these were recovered from 353 specimens collected from 247 patients. The greatest number of isolates of mycobacteria (289, or 80% of the 362 isolates) was recovered with the BACTEC MGIT 960, followed by the BACTEC 460 TB (271, or 75%) and solid media (250, or 69%). From all culture systems a total of 132 isolates of Mycobacterium tuberculosiscomplex were recovered. The greatest number of isolates of M. tuberculosis complex was recovered when liquid medium was combined with conventional solid media; the number recovered with BACTEC 460 TB plus solid media was 128 (97%), that recovered with BACTEC MGIT 960 plus solid media was 121 (92%), that recovered with BACTEC 460 TB was 119 (90%) and that recovered with all solid media combined was 105 (79%). The recovery with BACTEC MGIT 960 alone was 102 (77%). The mean times to detection (TTD) for M. tuberculosis complex were 14.4 days for BACTEC MGIT, 15.2 days for BACTEC 460 TB, and 24.1 days for solid media. The numbers of isolates of Mycobacterium avium complex (MAC) recovered were 172 (100%) for all systems, 147 (85%) for BACTEC MGIT 960, 123 (72%) for BACTEC 460 TB, and 106 (62%) for all solid media combined. The TTD for MAC in each system were 10.0 days for BACTEC MGIT 960, 10.4 days for BACTEC 460 TB, and 25.9 days for solid media. Breakthrough contamination rates (percentages of isolates) for each of the systems were 8.1% for BACTEC MGIT 960, 4.9% for BACTEC 460 TB, and 21.1% for all solid media combined.

Mono-and biexponential killing curves for vancomycin over a 2-to 50-p,g/ml concentration range were generated for 11 Staphylococcus aureus isolates and 12 coagulase-negative Staphylococcus species in the logarithmic phase of growth. Nonlinear least-squares regression of the initial growth rate and disappearance were not significantly different for lower or higher concentrations of vancomycin in broth.Interest in vancomycin has increased because of 3-lactam antibiotic resistance among gram-positive cocci. Recently, the disposition of and pharmacodynamic killing by vancomycin and teicoplanin have been characterized (2-5, 13, 17). Several predictive methods that propose the modification of aminoglycoside dosing models by using desired peak concentrations of 30 to 40 ,ug/ml and trough concentrations of 5 to 10 ,ug/ml have been proposed (15,22). Geraci and Hermans (11) recommended trough vancomycin concentrations of 8 to 10 ,ug/ml on the basis of an average MIC of 1 to 1.25 ,ug/ml, which has increased slightly since 1976. Carret et al.(7) described a unique application of a nonlinear regression program that is normally used for pharmacokinetic analysis. Killing curves generated from a range of ciprofloxacin concentrations were tested against Escherichia coli. The killing curves demonstrated a concentration-dependent lag phase in bacterial killing and parallel rate constants (7). The investigators (7) studied the in vitro killing of staphylococci by vancomycin in broth over a concentration range to determine whether any change in killing rate could be attributed to differences in the vancomycin concentration (1, 3, 8-10, 12, 19, 23, 24, 26, 27).Vancomycin hydrochloride (Eli Lilly & Company) was prepared as a stock solution (10 mg/ml) in phosphatebuffered saline (pH 6.0) and was frozen at -70°C in 1.5-ml aliquots. Prepared solutions were assayed every 1 to 2 weeks by using a commercial fluorescence polarization immunoassay (TDx; Abbott Laboratories, North Chicago, Ill.) (25).Staphylococcus aureus isolates were obtained from blood cultures (n = 8), respiratory secretions (n = 2), and a wound abscess (n = 1). Coagulase-negative Staphylococcus isolates were obtained from blood cultures (n = 3), abscess wound cultures (n = 4), pleural fluid culture (n = 1), urine culture (n = 1), coagulase-positive Staphylococcus saprophyticus (n = 1), and Staphylococcus epidermidis ATCC 12228 (n = 1). 50, 30,20,10,8,6,4, and 2 ,ug/ml. A growth control with no added antibiotic solution was also used. All tubes were run in duplicate.Time-kill curves were made for tubes containing bacterial suspensions and the various vancomycin concentrations incubated at 35°C. At 1, 2, 3, 4, 5, 6, 8, and 24 h postexposure, aliquots were diluted in tryptic soy broth (Difco) and plated onto blood agar plates (BBL, Cockeysville, Md.). Aliquots were made by using a clean sterile pipette tip for each tube. The number of CFU per milliliter was determined after 24 h of incubation.Killing curves were analyzed by using PKCALC, a BASIC program containing the curve...