"Shortcuts" Causing Bias in Radiology Artificial Intelligence: Causes, Evaluation, and Mitigation

Despite the expert-level performance of artificial intelligence (AI) models for various medical imaging tasks, real-world performance failures with disparate outputs for various subgroups limit the usefulness of AI in improving patients' lives. Many definitions of fairness have been proposed, with discussions of various tensions that arise in the choice of an appropriate metric to use to evaluate bias; for example, should one aim for individual or group fairness? One central observation is that AI models apply "shortcut learning" whereby spurious features (such as chest tubes and portable radiographic markers on intensive care unit chest radiography) on medical images are used for prediction instead of identifying true pathology. Moreover, AI has been shown to have a remarkable ability to detect protected attributes of age, sex, and race, while the same models demonstrate bias against historically underserved subgroups of age, sex, and race in disease diagnosis. Therefore, an AI model may take shortcut predictions from these correlations and subsequently generate an outcome that is biased toward certain subgroups even when protected attributes are not explicitly used as inputs into the model. As a result, these subgroups became nonprivileged subgroups. In this review, the authors discuss the various types of bias from shortcut learning that may occur at different phases of AI model development, including data bias, modeling bias, and inference bias. The authors thereafter summarize various tool kits that can be used to evaluate and mitigate bias and note that these have largely been applied to nonmedical domains and require more evaluation for medical AI. The authors then summarize current techniques for mitigating bias from preprocessing (data-centric solutions) and during model development (computational solutions) and postprocessing (recalibration of learning). Ongoing legal changes where the use of a biased model will be penalized highlight the necessity of understanding, detecting, and mitigating biases from shortcut learning and will require diverse research teams looking at the whole AI pipeline.

Evaluation of federated learning variations for COVID-19 diagnosis using chest radiographs from 42 US and European hospitals

OBJECTIVE

Federated learning (FL) allows multiple distributed data holders to collaboratively learn a shared model without data sharing. However, individual health system data are heterogeneous. "Personalized" FL variations have been developed to counter data heterogeneity, but few have been evaluated using real-world healthcare data. The purpose of this study is to investigate the performance of a single-site versus a 3-client federated model using a previously described Coronavirus Disease 19 (COVID-19) diagnostic model. Additionally, to investigate the effect of system heterogeneity, we evaluate the performance of 4 FL variations.

MATERIALS AND METHODS

We leverage a FL healthcare collaborative including data from 5 international healthcare systems (US and Europe) encompassing 42 hospitals. We implemented a COVID-19 computer vision diagnosis system using the Federated Averaging (FedAvg) algorithm implemented on Clara Train SDK 4.0. To study the effect of data heterogeneity, training data was pooled from 3 systems locally and federation was simulated. We compared a centralized/pooled model, versus FedAvg, and 3 personalized FL variations (FedProx, FedBN, and FedAMP).

RESULTS

We observed comparable model performance with respect to internal validation (local model: AUROC 0.94 vs FedAvg: 0.95, P = .5) and improved model generalizability with the FedAvg model (P < .05). When investigating the effects of model heterogeneity, we observed poor performance with FedAvg on internal validation as compared to personalized FL algorithms. FedAvg did have improved generalizability compared to personalized FL algorithms. On average, FedBN had the best rank performance on internal and external validation.

CONCLUSION

FedAvg can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity. Personalized FL may offer an opportunity to develop both internal and externally validated algorithms.

Performance of a Chest Radiograph AI Diagnostic Tool for COVID-19: A Prospective Observational Study

PURPOSE

To conduct a prospective observational study across 12 U.S. hospitals to evaluate real-time performance of an interpretable artificial intelligence (AI) model to detect COVID-19 on chest radiographs.

MATERIALS AND METHODS

A total of 95 363 chest radiographs were included in model training, external validation, and real-time validation. The model was deployed as a clinical decision support system, and performance was prospectively evaluated. There were 5335 total real-time predictions and a COVID-19 prevalence of 4.8% (258 of 5335). Model performance was assessed with use of receiver operating characteristic analysis, precision-recall curves, and F1 score. Logistic regression was used to evaluate the association of race and sex with AI model diagnostic accuracy. To compare model accuracy with the performance of board-certified radiologists, a third dataset of 1638 images was read independently by two radiologists.

RESULTS

Participants positive for COVID-19 had higher COVID-19 diagnostic scores than participants negative for COVID-19 (median, 0.1 [IQR, 0.0-0.8] vs 0.0 [IQR, 0.0-0.1], respectively; P < .001). Real-time model performance was unchanged over 19 weeks of implementation (area under the receiver operating characteristic curve, 0.70; 95% CI: 0.66, 0.73). Model sensitivity was higher in men than women (P = .01), whereas model specificity was higher in women (P = .001). Sensitivity was higher for Asian (P = .002) and Black (P = .046) participants compared with White participants. The COVID-19 AI diagnostic system had worse accuracy (63.5% correct) compared with radiologist predictions (radiologist 1 = 67.8% correct, radiologist 2 = 68.6% correct; McNemar P < .001 for both).

CONCLUSION

AI-based tools have not yet reached full diagnostic potential for COVID-19 and underperform compared with radiologist prediction.Keywords: Diagnosis, Classification, Application Domain, Infection, Lung Supplemental material is available for this article.. © RSNA, 2022.

AI recognition of patient race in medical imaging: a modelling study

BACKGROUND

Previous studies in medical imaging have shown disparate abilities of artificial intelligence (AI) to detect a person's race, yet there is no known correlation for race on medical imaging that would be obvious to human experts when interpreting the images. We aimed to conduct a comprehensive evaluation of the ability of AI to recognise a patient's racial identity from medical images.

METHODS

Using private (Emory CXR, Emory Chest CT, Emory Cervical Spine, and Emory Mammogram) and public (MIMIC-CXR, CheXpert, National Lung Cancer Screening Trial, RSNA Pulmonary Embolism CT, and Digital Hand Atlas) datasets, we evaluated, first, performance quantification of deep learning models in detecting race from medical images, including the ability of these models to generalise to external environments and across multiple imaging modalities. Second, we assessed possible confounding of anatomic and phenotypic population features by assessing the ability of these hypothesised confounders to detect race in isolation using regression models, and by re-evaluating the deep learning models by testing them on datasets stratified by these hypothesised confounding variables. Last, by exploring the effect of image corruptions on model performance, we investigated the underlying mechanism by which AI models can recognise race.

FINDINGS

In our study, we show that standard AI deep learning models can be trained to predict race from medical images with high performance across multiple imaging modalities, which was sustained under external validation conditions (x-ray imaging [area under the receiver operating characteristics curve (AUC) range 0·91-0·99], CT chest imaging [0·87-0·96], and mammography [0·81]). We also showed that this detection is not due to proxies or imaging-related surrogate covariates for race (eg, performance of possible confounders: body-mass index [AUC 0·55], disease distribution [0·61], and breast density [0·61]). Finally, we provide evidence to show that the ability of AI deep learning models persisted over all anatomical regions and frequency spectrums of the images, suggesting the efforts to control this behaviour when it is undesirable will be challenging and demand further study.

INTERPRETATION

The results from our study emphasise that the ability of AI deep learning models to predict self-reported race is itself not the issue of importance. However, our finding that AI can accurately predict self-reported race, even from corrupted, cropped, and noised medical images, often when clinical experts cannot, creates an enormous risk for all model deployments in medical imaging.

FUNDING

National Institute of Biomedical Imaging and Bioengineering, MIDRC grant of National Institutes of Health, US National Science Foundation, National Library of Medicine of the National Institutes of Health, and Taiwan Ministry of Science and Technology.

Enhancing cancer prevention and survivorship care with a videoconferencing model for continuing education: a mixed-methods study to identify barriers and incentives to participation

Objective

To enhance cancer prevention and survivorship care by local health care providers, a school of public health introduced an innovative telelearning continuing education program using the Extension for Community Healthcare Outcomes (ECHO) model. In ECHO’s hub and spoke structure, synchronous videoconferencing connects frontline health professionals at various locations (“spokes”) with experts at the facilitation center (“hub”). Sessions include experts’ didactic presentations and case discussions led by spoke site participants. The objective of this study was to gain a better understanding of the reasons individuals choose or decline to participate in the Cancer ECHO program and to identify incentives and barriers to doing so.

Materials and methods

Study participants were recruited from the hub team, spoke site participants, and providers who attended another ECHO program but not this one. Participants chose to take a survey or be interviewed. The Consolidated Framework for Implementation Research guided qualitative data coding and analysis.

Results

We conducted 22 semistructured interviews and collected 30 surveys. Incentives identified included the program’s high-quality design, supportive learning climate, and access to information. Barriers included a lack of external incentives to participate and limited time available. Participants wanted more adaptability in program timing to fit providers’ busy schedules.

Conclusion

Although the merits of the Cancer ECHO program were widely acknowledged, adaptations to facilitate participation and emphasize the program’s benefits may help overcome barriers to attending. As the number of telelearning programs grows, the results of this study point to ways to expand participation and spread health benefits more widely.

Cancer remains the second leading cause of death in the United States, and new methods are needed to support health care professionals in providing both prevention and survivorship care for their patients. To that end, a school of public health introduced an innovative telelearning program to use technology to link frontline health professionals at their locations with experts at the school for regularly scheduled learning sessions. These sessions on a range of cancer prevention and survivorship care topics take place twice a month and feature presentation of current information and discussion of cases. After the program’s pilot year, this study sought input from both participants and nonparticipants to identify incentives and barriers to participation. Input was collected via interviews and a survey. Incentives to participate included the program’s high-quality design, supportive learning climate, and access to information. Barriers included a lack of external incentives to participate and limited time available. Participants also wanted more adaptability in program timing to fit busy health care provider schedules. The results will be used to expand and improve the program in the future with the aim of enhancing providers’ use of cancer prevention strategies and care for survivors.

A Prospective Observational Study to Investigate Performance of a Chest X-ray Artificial Intelligence Diagnostic Support Tool Across 12 U.S. Hospitals

Importance

An artificial intelligence (AI)-based model to predict COVID-19 likelihood from chest x-ray (CXR) findings can serve as an important adjunct to accelerate immediate clinical decision making and improve clinical decision making. Despite significant efforts, many limitations and biases exist in previously developed AI diagnostic models for COVID-19. Utilizing a large set of local and international CXR images, we developed an AI model with high performance on temporal and external validation.

Objective

Investigate real-time performance of an AI-enabled COVID-19 diagnostic support system across a 12-hospital system.

Design

Prospective observational study.

Setting

Labeled frontal CXR images (samples of COVID-19 and non-COVID-19) from the M Health Fairview (Minnesota, USA), Valencian Region Medical ImageBank (Spain), MIMIC-CXR, Open-I 2013 Chest X-ray Collection, GitHub COVID-19 Image Data Collection (International), Indiana University (Indiana, USA), and Emory University (Georgia, USA).

Participants

Internal (training, temporal, and real-time validation): 51,592 CXRs; Public: 27,424 CXRs; External (Indiana University): 10,002 CXRs; External (Emory University): 2002 CXRs.

Main Outcome and Measure

Model performance assessed via receiver operating characteristic (ROC), Precision-Recall curves, and F1 score.

Results

Patients that were COVID-19 positive had significantly higher COVID-19 Diagnostic Scores (median .1 [IQR: 0.0-0.8] vs median 0.0 [IQR: 0.0-0.1], p < 0.001) than patients that were COVID-19 negative. Pre-implementation the AI-model performed well on temporal validation (AUROC 0.8) and external validation (AUROC 0.76 at Indiana U, AUROC 0.72 at Emory U). The model was noted to have unrealistic performance (AUROC > 0.95) using publicly available databases. Real-time model performance was unchanged over 19 weeks of implementation (AUROC 0.70). On subgroup analysis, the model had improved discrimination for patients with "severe" as compared to "mild or moderate" disease, p < 0.001. Model performance was highest in Asians and lowest in whites and similar between males and females.

Conclusions and Relevance

AI-based diagnostic tools may serve as an adjunct, but not replacement, for clinical decision support of COVID-19 diagnosis, which largely hinges on exposure history, signs, and symptoms. While AI-based tools have not yet reached full diagnostic potential in COVID-19, they may still offer valuable information to clinicians taken into consideration along with clinical signs and symptoms.

Just in Time Radiology Decision Support Using Real-time Data Feeds

Ready access to relevant real-time information in medical imaging offers several potential benefits. Knowing both when important information will be available and that important information is available can facilitate optimization of workflow and management of time. Unexpected findings, as well as deficiencies in reporting and documentation, can be immediately managed. Herein, we present our experience developing and implementing a real-time web-centric dashboard system for radiologists, clinicians, and support staff. The dashboards are driven by multi-sourced HL7 message streams that are monitored, analyzed, aggregated, and transformed into multiple real-time displays to improve operations within our department. We call this framework Pipeline. Ruby on Rails, JavaScript, HTML, and SQL serve as the foundations of the Pipeline application. HL7 messages are processed in real-time by a Mirth interface engine which posts exam data into SQL. Users utilize web browsers to visit the Ruby on Rails-based dashboards on any device connected to our hospital network. The dashboards will automatically refresh every 30 seconds using JavaScript. The Pipeline application has been well received by clinicians and radiologists.

Characterization of the nasal, sublingual, and oropharyngeal mucosa microbiota in cleft lip and palate individuals before and after surgical repair

OBJECTIVE

To delineate inherent differences in the microbial milieu in cleft palate patients compared with cleft lip patients and to document changes in microbial flora before and after cleft lip and palate repair.

DESIGN

A prospective study of preoperative and postoperative culture results from the nasal, sublingual, and oropharyngeal surfaces of patients undergoing primary cleft lip repair and palate closure.

SETTING

Shriners Hospitals for Children, Galveston, Texas, and University of Texas Medical Branch, Galveston, Texas.

PATIENTS

Seventy-nine patients were included in a 3-year period. Ten patients with isolated cleft lip underwent primary lip repair. Twenty-five patients with cleft lip and palate underwent primary lip repair, and 44 patients underwent palatoplasty.

RESULTS

Cleft palate patients had a significantly higher rate of colonization by staphylococcal species, but not methicillin-resistant Staphylococcus aureus , when compared to cleft lip patients (p=.0298; chi-square test). Closure of the palatal cleft coincided with significant decline in the prevalence of Klebsiella and Enterobacter species (p<.05; McNemar test). The only major complication, palatal dehiscence, was believed to be directly related to infection with group A beta-hemolytic streptococci.

CONCLUSIONS

Despite a high prevalence of potential pathogenic and enteric flora preoperatively in primary palate repair, postoperative wound infection is rare in the prospective study population. However, the presence of beta-hemolytic streptococci was associated with a higher risk of repair dehiscence; therefore, screening for Streptococci prior to surgery should be performed routinely.

Phase II studies of nebulised Arikace in CF patients with Pseudomonas aeruginosa infection

Rationale

Arikace is a liposomal amikacin preparation for aerosol delivery with potent Pseudomonas aeruginosa killing and prolonged lung deposition.

Objectives

To examine the safety and efficacy of 28 days of once-daily Arikace in cystic fibrosis (CF) patients chronically infected with P aeruginosa.

Methods

105 subjects were evaluated in double-blind, placebo-controlled studies. Subjects were randomised to once-daily Arikace (70, 140, 280 and 560 mg; n=7, 5, 21 and 36 subjects) or placebo (n=36) for 28 days. Primary outcomes included safety and tolerability. Secondary outcomes included lung function (forced expiratory volume at one second (FEV₁)), P aeruginosa density in sputum, and the Cystic Fibrosis Quality of Life Questionnaire—Revised (CFQ-R).

Results

The adverse event profile was similar among Arikace and placebo subjects. The relative change in FEV₁ was higher in the 560 mg dose group at day 28 (p=0.033) and at day 56 (28 days post-treatment, 0.093L±0.203 vs −0.032L±0.119; p=0.003) versus placebo. Sputum P aeruginosa density decreased >1 log in the 560 mg group versus placebo (days 14, 28 and 35; p=0.021). The Respiratory Domain of the CFQ-R increased by the Minimal Clinically Important Difference (MCID) in 67% of Arikace subjects (560 mg) versus 36% of placebo (p=0.006), and correlated with FEV₁ improvements at days 14, 28 and 42 (p<0.05). An open-label extension (560 mg Arikace) for 28 days followed by 56 days off over six cycles confirmed durable improvements in lung function and sputum P aeruginosa density (n=49).

Conclusions

Once-daily Arikace demonstrated acute tolerability, safety, biologic activity and efficacy in patients with CF with P aeruginosa infection.

Evolution of Pseudomonas aeruginosa virulence in infected patients revealed in a Dictyostelium discoideum host model

Pseudomonas aeruginosa can cause acute lung infections in intubated patients or chronic infections in patients with cystic fibrosis (CF). In both situations, P. aeruginosa accumulates specific mutations, in particular in the lasR quorum-sensing regulator gene. Using a Dictyostelium discoideum amoeba model, we assessed whether these mutations affect bacterial virulence. Among a collection of clinical isolates from 16 CF patients, initial isolates were fully virulent in 15 patients, but for late isolates collected several years later, virulence was decreased in eight patients. No significant correlation between genetic inactivation of lasR and decreased virulence was observed. Among strains isolated from ten colonized intubated patients, all initial isolates were fully virulent. Despite the accumulation of lasR-inactivating mutations in strains collected over a 3-week period, no decrease in virulence was observed in eight of 10 patients. In one intubated patient, the virulent initial strain was replaced a few days later with a different, less virulent, strain. We observed a gradual decrease in bacterial virulence in only one intubated patient. We conclude that adaptation of P. aeruginosa to chronically infected CF patients can lead to a slow and gradual loss of virulence, as measured in a Dictyostelium model system. However, loss of virulence is not caused predominantly by mutations in lasR. During short-term colonization of intubated patients for up to 20 days, a decrease in virulence was exceptional, despite the accumulation of lasR mutations.

Genetic adaptation of Pseudomonas aeruginosa to the airways of cystic fibrosis patients is catalyzed by hypermutation

In previous work (E. E. Smith, D. G. Buckley, Z. Wu, C. Saenphimmachack, L. R. Hoffman, D. A. D'Argenio, S. I. Miller, B. W. Ramsey, D. P. Speert, S. M. Moskowitz, J. L. Burns, R. Kaul, and M. V. Olson, Proc. Natl. Acad. Sci. USA 103:8487-8492, 2006) it was shown that Pseudomonas aeruginosa undergoes intense genetic adaptation during chronic respiratory infection (CRI) in cystic fibrosis (CF) patients. We used the same collection of isolates to explore the role of hypermutation in this process, since one of the hallmarks of CRI is the high prevalence of DNA mismatch repair (MMR) system-deficient mutator strains. The presence of mutations in 34 genes (many of them positively linked to adaptation in CF patients) in the study collection of 90 P. aeruginosa isolates obtained longitudinally from 29 CF patients was not homogeneous; on the contrary, mutations were significantly concentrated in the mutator lineages, which represented 17% of the isolates (87% MMR deficient). While sequential nonmutator lineages acquired a median of only 0.25 mutation per year of infection, mutator lineages accumulated more than 3 mutations per year. On the whole-genome scale, data for the first fully sequenced late CF isolate, which was also shown to be an MMR-deficient mutator, also support these findings. Moreover, for the first time the predicted amplification of mutator populations due to hitchhiking with adaptive mutations in the course of natural human infections is clearly documented. Interestingly, increased accumulation of mutations in mutator lineages was not a consequence of overrepresentation of mutations in genes involved in antimicrobial resistance, the only adaptive trait linked so far to hypermutation in CF patients, demonstrating that hypermutation also plays a major role in P. aeruginosa genome evolution and adaptation during CRI.

Noninvasive body contouring by focused ultrasound: safety and efficacy of the Contour I device in a multicenter, controlled, clinical study

BACKGROUND

The removal of unwanted body fat using a noninvasive technique is desirable to patients and physicians. The authors describe a controlled, multicenter, clinical trial assessing the safety and efficacy of a focused therapeutic ultrasound device for noninvasive body contouring.

METHODS

Eligible healthy adult subjects were enrolled to the experimental group or the control group at five sites. The experimental group received one treatment with the Contour I device (UltraShape Ltd., Tel Aviv, Israel) in the abdomen, thighs, or flanks and were evaluated over a 12-week period. Efficacy outcomes were reduction of circumference and fat thickness. Circumference reduction was compared with the untreated group and with an untreated area (thigh) within the treated group. Safety monitoring included laboratory testing (including serum lipids), pulse oximetry, and liver ultrasound.

RESULTS

One hundred sixty-four subjects participated in the study (137 subjects in the experimental group and 27 in the control, untreated group). A single Contour I treatment was safe and well tolerated and produced a mean reduction of approximately 2 cm in treatment area circumference and approximately 2.9 mm in skin fat thickness. The majority of the effect was achieved within 2 weeks and was sustained at 12 weeks. No clinically significant changes in the measured safety parameters were recorded. Seven adverse events were reported, all of which were anticipated, mild, and resolved within the study period.

CONCLUSION

The Contour I device provides a safe and effective noninvasive technology for body contouring.

Covariates of tooth-brushing frequency in low-income African Americans from grades 5 to 8

PURPOSE

The purpose of this study was to examine tooth-brushing frequency in 575 urban and nearby suburban African American children as part of a comprehensive risk-reduction study for students at high risk for violence, drugs, school delinquency, and unsafe sexual behaviors to determine which covariates predicted tooth-brushing frequency.

METHODS

Students were surveyed 5 times, from the beginning of grade 5 and the end of each year through grade 8, and parents were surveyed at the beginning of grade 5. Peer influence, importance of being liked, self-esteem, attitudes towards tooth-brushing, oral health knowledge, self-efficacy, parental attitudes, and other covariates were examined for the ability to predict self-reporting of tooth-brushing frequency.

RESULTS

In the fifth grade, peer influence, the importance of being liked, and physical self-esteem were the significant predictors, and peer influence continued to predict tooth-brushing in the eighth grade. Oral health knowledge and parental influence were not significant.

CONCLUSION

Peer influence is an important factor in tooth-brushing behavior in metropolitan African American preadolescent children.

Insulin-like growth factor ligands, receptors, and binding proteins in cancer

This review aims to summarize experimental evidence supporting the role of the insulin-like growth factor (IGF) signalling system in the progression, maintenance, and treatment of cancer. These data implicate the IGF system as an important modifier of cancer cell proliferation, survival, growth, and treatment sensitivity. The role of the IGF system in cancer should be examined in the context of the extra-cellular and intra-cellular signalling networks, in particular: phosphatidylinositol 3-kinase (PI3K), protein kinase B (Akt/PKB), mammalian target of rapamycin (mTOR), and forkhead transcription factors (FOXO). This review highlights evidence derived from molecular structure and functional genetics with respect to how the extra-cellular components of the IGF system function normally, and their subsequent modifications in cancer. The therapeutic relevance of the research evidence described is also addressed, as the challenge is to apply this knowledge to human health.

Impairments to wound healing

Impaired wound healing is a complication faced by all physicians, regardless of their field of practice. Plastic surgeons are frequently called on to help treat patients who fail to heal properly. Therefore, plastic surgeons must be well versed in the intrinsic and extrinsic factors that can impair wound healing, such as nutrition, drugs, radiation, smoking, and hypoxia. Only by limiting detrimental factors can wound healing progress in a beneficial fashion.

Viscosity effect on the structural compactness of latex flocs formed under weak depletion attractions

Dilute aqueous dispersions of colloidal polystyrene latex spheres were flocculated by adding a nonadsorbing polymer sample, poly(acrylic acid). The structural compactness of the flocs thus formed was characterized in terms of their mass fractal dimension using the small-angle static light scattering technique. It was found that with low poly(acrylic acid) concentrations and thus weak depletion attraction forces, the dispersion medium viscosity had a marked effect on the floc structure. An increase in the viscosity led to formation of denser flocs. This was revealed in three sets of depletion flocculation experiments: (a) adjusting the background electrolyte concentration at a fixed level of poly(acrylic acid), (b) using water and 30% (w/w) glycerol as the respective solvents, and (c) inducing latex flocculation with two poly(acrylic acids) of different molecular weights at the respective critical polyacid concentrations. Direct force measurements were made with atomic force microscopy to isolate the influence of viscosity on floc structure from that of interparticle interaction energies. We conclude that the formation of denser flocs with increasing medium viscosity can be attributed to the reduced diffusivity of particles in the solution. The latter resulted in an enhanced rate of floc restructuring (through relaxation of attached particles) relative to floc growth.

Early pulmonary infection, inflammation, and clinical outcomes in infants with cystic fibrosis

A thorough understanding of the early natural history of cystic fibrosis (CF) lung disease is critical for the development of effective interventions in the youngest patients. We assessed the evolution of pulmonary infection, inflammation, and clinical course among 40 infants over a 2-year period through annual bronchoalveolar lavage (BAL) for culture and measurements of pro- and anti-inflammatory cytokines, semiannual infant pulmonary function testing, and quarterly clinical evaluations. Both the prevalence of CF pathogens and their density in BAL fluid increased with age. Infants had neutrophilic lower airway inflammation and elevated IL-8 concentrations independent of whether CF pathogens were recovered. Total leukocyte and neutrophil densities and IL-8 concentrations increased with density of CF pathogens in BAL fluid, whether the isolated organism was P. aeruginosa or another pathogen. IL-10 concentrations were similar in CF subjects and non-CF historical controls. Infants generally had suboptimal growth (low weight and height percentiles) and obstructive lung disease (decreased expiratory flows and air trapping). Subjects from whom CF pathogens were isolated at > 10(5) cfu/mL had the worst air trapping and lowest Brasfield chest X-ray scores. Our findings provide a foundation for future studies of early intervention in CF lung disease, including antimicrobial and anti-inflammatory therapy.

Cell survival and proliferation are modified by insulin-like growth factor 2 between days 9 and 10 of mouse gestation

The size of mammalian species involves the interaction of multiple genetic modifiers that control the timing and extent of growth mechanisms. Disruption of the paternal allele of the imprinted embryonic gene coding for insulin-like growth factor 2 (IGF2, Igf2+m/−p), results in viable mice that are 60% the weight of wild-type littermates. Differences in weight are first detected at embryonic day (E) 11, and the growth deficit is maintained throughout life. We report the mechanisms that account for this unusual phenotype. In order to quantify growth, we used novel methods to generate single cell suspensions of post-implantation mouse embryos. We were then able to quantify cell number, cell proliferation and cell death between E8.5 and E11.5 using flow cytometry. Determination of total embryo cell number also allowed us to time litters by a method other than by plugging. Wild-type and Igf2+m/−p embryos accumulated similar total cell numbers up to E9.25, but cell number began to diverge by around E9.5, with significant differences by E11 (75% of wild type). A relative increase in pyknotic nuclei, sub-GI cytometry counts and caspase activity, all indicative of cell death, occurred in Igf2+m/−p embryos at E9.25, reverting to wild-type levels by E9.75. This was followed at E9.75 by a significant reduction in the proportion of cells in S phase, quantified by S-phase cytometry counts and BrdU labelling. No significant differences in cell size were detected. We conclude that the majority of the cell number differences between wild-type and Igf2+m/−p mice can be accounted for by modification of cell survival and proliferation during the period (E9 to E10) of post-implantation development.

Serum and lower respiratory tract drug concentrations after tobramycin inhalation in young children with cystic fibrosis

OBJECTIVES

To assess the serum and lower respiratory tract tobramycin concentrations (C(T)) produced by a single dose of tobramycin for inhalation delivered by a nebulizer and a compressor in patients with cystic fibrosis (CF) 6 months to 6 years of age.

STUDY DESIGN

We performed a dose escalation study of serum C(T) measured before and 0.5, 1, 2, and 4 hours after a single dose of inhaled tobramycin, either 180 mg (10 patients) or 300 mg (19 patients). In a separate group of 12 patients, epithelial lining fluid (ELF) C(T) was measured by bronchoalveolar lavage 30 to 45 minutes after a 300-mg dose.

RESULTS

A 180-mg dose of inhaled tobramycin produced a mean peak serum C(T) of 0.5 microg/mL (SD 0.4; range, <0.2 to 1.4 microg/mL). A 300-mg dose produced a mean peak serum C(T) of 0.6 microg/mL (SD 0.5; range, <0.2 to 1.2 microg/mL). These peak values are well below the accepted maximum trough concentration with parenteral dosing (2 microg/mL). The target ELF C(T) was 20 microg/mL, 10-fold greater than the minimal inhibitory concentration for most Pseudomonas aeruginosa isolates from very young patients with CF (2 microg/mL). Mean ELF C(T) was 90 microg/mL (SD 54; range, 16 to 204 microg/mL) and exceeded the target concentration in 11 patients.

CONCLUSION

In patients with CF ages 6 months to 6 years, a single 300-mg dose of inhaled tobramycin appears to produce safe peak serum concentrations and drug concentrations in the bactericidal range in the lower respiratory tract.

Use of random amplified polymorphic DNA PCR to examine epidemiology of Stenotrophomonas maltophilia and Achromobacter (Alcaligenes) xylosoxidans from patients with cystic fibrosis

Stenotrophomonas maltophilia and Achromobacter (Alcaligenes) xylosoxidans have been increasingly recognized as a cause of respiratory tract colonization in cystic fibrosis (CF). Although both organisms have been associated with progressive deterioration of pulmonary function, demonstration of causality is lacking. To examine the molecular epidemiology of S. maltophilia and A. xylosoxidans in CF, isolates from patients monitored for up to 2 years were fingerprinted using a PCR-based randomly amplified polymorphic DNA (RAPD-PCR) method. Sixty-one of 69 CF centers screened had 183 S. maltophilia culture-positive patients, and 46 centers had 92 A. xylosoxidans-positive patients. At least one isolate from each patient was genotyped, and patients with > or =10 positive cultures (12 S. maltophilia cultures, 15 A. xylosoxidans cultures) had serial isolates genotyped. In addition, centers with multiple culture-positive patients were examined for evidence of shared clones. There were no instances of shared genotypes among different CF centers. Some patients demonstrated isolates with a single genotype throughout the observation period, and others had intervening or sequential genotypes. At the six centers with multiple S. maltophilia culture-positive patients and the seven centers with multiple A. xylosoxidans-positive patients, there were three and five instances of shared genotypes, respectively. The majority of shared isolates were from pairs who were siblings or otherwise epidemiologically linked. These findings suggest RAPD-PCR typing can distinguish unique CF isolates of S. maltophilia and A. xylosoxidans, person-to-person transmission may occur, there are not a small number of clones infecting CF airways, and patients with long-term colonization may either have a persistent organism or may acquire additional organisms over time.

A view through the clouds of imprinting

The purpose of this review is to examine whether our current knowledge of the higher order control of gene expression and nuclear organization can help us understand the mechanisms of genomic imprinting. Imprinting involves the inheritance of a silenced allele of a gene through either a paternal or maternal germline. We have approached the problem of imprinting using a model based on the dynamic attachment of chromatin loops to immobilized RNA polymerases and control elements. We have combined the information from different experimental approaches, examining primarily the IGF2-H19 locus, in an attempt to simplify the complexity of the imprinting data that has accumulated. It is hoped that a unified model may generate predictions amenable to experimental testing and contribute to the interpretation of future experiments.

A rare continual flowering strategy and its influence on offspring quality in a gynodioecious plant

The majority of angiosperms have a single annual bout of reproduction; species that flower continually throughout the year are rare. Ochradenus baccatus is a gynodioecious, desert shrub whose principal flowering period is associated with the winter rains, although large individuals also remain in flower during the hot, dry summer. The goal of this study was to examine the reproductive consequences of continual flowering in a large population of O. baccatus in Israel. Over the two years of this study, 60% of individuals flowered continuously. The number of fruit and seeds per fruit were greater in winter. Winter seeds were ∼12% heavier than summer seeds and had significantly higher germination rates (80 vs. 60%, respectively). Although summer seeds were smaller and less likely to germinate, we propose that the benefit derived from their production lies in their ability to capitalize on the first winter rains. These early rain events provide a head start on establishment and growth in the hostile desert environment. Plants that delay reproduction until the onset of rains risk having their offspring face the dry conditions of spring and summer.

Young at heart: understanding the unique psychosocial adjustment of young implantable cardioverter defibrillator recipients

This article reviews the data related to psychosocial adjustment of young ICD recipients, postulates theories to explain potential adjustment difficulties to ICD therapy experienced by younger recipients, and suggests clinical management techniques for addressing the unique psychosocial concerns of young ICD recipients. Studies of young ICD recipients suggest that a wide range of psychosocial adjustment issues are prominent in the post-ICD implantation period and that the issues may be different from older ICD recipients. The disability-stress-coping model and the transactional-stress-coping model are postulated as explanations for the unique adjustment concerns of children and adolescents with ICDs. Social comparison theory is also applied to the concerns of young adults with ICDs such that they often lack same age peers to compare experiences with cardiac difficulties. Brief, clinic-based interventions by health care providers, like a screening and referral heuristic and an "ICD Buddy" system, are suggested to increase effective coping and decrease social isolation for young ICD recipients.

Comparison of two commercial systems (Vitek and MicroScan-WalkAway) for antimicrobial susceptibility testing of Pseudomonas aeruginosa isolates from cystic fibrosis patients

Antimicrobial susceptibility testing of cystic fibrosis (CF) isolates of Pseudomonas aeruginosa is difficult because the organisms are often mucoid and slow-growing. This study of 498 CF strains examined the correlation of results derived from two commonly used commercial systems (Vitek, MicroScan-WalkAway) with a reference method for 10 antimicrobials. Correlation to reference results was unacceptably low for all agents and both commercial systems had a high rate of very major (false-susceptible) errors. Although mucoid strains produced a 4.8% greater intermethod error, it was not markedly different than non-mucoid strains for the Vitek System. Overall, these tested commercial systems performed poorly for CF isolates in contrast to earlier reported, high correlations with the reference methods (broth microdilution frozen panels and agar dilution) of the National Committee for Clinical Laboratory Standards, the standardized disk diffusion test, and the Etest (AB BIODISK, Solna, Sweden).

Sputum induction as a research tool for sampling the airways of subjects with cystic fibrosis

BACKGROUND

Sputum induction (SI) has proved to be a reliable non-invasive tool for sampling inflammatory airway contents in asthma, with distinct advantages over collection of expectorated sputum (ES) and bronchoalveolar lavage (BAL). A study was undertaken to evaluate the safety of SI and to assess if it might be an equally valuable outcome tool in patients with cystic fibrosis (CF).

METHODS

The safety of the procedure was examined and sample volume, cell counts, cytokine concentrations, and bacterial culture results obtained by SI, spontaneous ES, and fibreoptic bronchoscopy were compared in 10 adults with CF.

RESULTS

SI was well tolerated and was preferred to BAL by all subjects. The mean (SE) sample volume obtained by SI was significantly greater than ES (6.74 (1.46) ml v 1.85 (0.33) ml, p = 0.005). There was no significant difference in the number of cells per ml of sample collected. There was a difference in the mean (SD) percentage of non-epithelial, non-squamous cells collected (67 (28)%, 86 (21)%, and 99 (1)% for ES, SI, and BAL, respectively). These percentage counts were different between ES and both SI and BAL (p=0.03 and p=0.006, respectively). Cell differential counts (excluding squamous cells) from all collection methods were similar (mean (SD) 84 (9)%, 87 (7)%, and 88 (11)% polymorphonuclear cells for ES, SI, and BAL, respectively). The concentrations of interleukin (IL)-8 and tumour necrosis factor (TNF)-alpha were the same in all three samples when corrected for dilution using urea concentration. The test specific detection rate for recovery of bacteriological pathogens was 79% for SI, 76% for ES, and 73% for BAL.

CONCLUSION

SI offers safety advantages over BAL and may be a more representative airway outcome measurement in patients with CF.

Longitudinal assessment of Pseudomonas aeruginosa in young children with cystic fibrosis

Pseudomonas aeruginosa lung infection is an important cause of morbidity and mortality in cystic fibrosis (CF). Longitudinal assessment of the phenotypic changes in P. aeruginosa isolated from young children with CF is lacking. This study investigated genotypic and phenotypic changes in P. aeruginosa from oropharynx (OP) and bronchoalveolar lavage fluid (BALF) in a cohort of 40 CF patients during the first 3 years of life; antibody response was also examined. A high degree of genotypic variability was identified, and each patient had unique genotypes. Early isolates had a phenotype distinct from those of usual CF isolates: generally nonmucoid and antibiotic susceptible. Genotype and phenotype correlated between OP and BALF isolates. As determined by culture, 72.5% of patients demonstrated P. aeruginosa during their first 3 years. On the basis of combined culture and serologic results, 97.5% of patients had evidence of infection by age 3 years, which suggests that P. aeruginosa infection occurs early in CF and may be intermittent or undetectable by culture.

Transcytosis of gastrointestinal epithelial cells by Escherichia coli K1

Escherichia coli K1 is an important neonatal pathogen that is usually transferred from maternal to infant gastrointestinal tract at the time of parturition. Approximately 20% of neonates are colonized, and a proportion of colonized infants goes on to have systemic infection. Entry into the bloodstream from the gastrointestinal tract is hypothesized to occur via epithelial cell invasion. Invasion of multiple epithelial cell lines was studied using gentamicin protection assays and transcytosis of polarized monolayers. Electron microscopy was used to confirm cellular invasion. Cell lines used include two human gastrointestinal lines, Caco-2 and T84; a human respiratory cell line, A549; a human laryngeal cell line, HEp-2; and a canine kidney cell line, MDCK. A virulent E. coli K1 strain, RS218, readily invaded HEp-2, A549, and T84 cell lines in gentamicin protection assays, but was less invasive into MDCK and Caco-2 cells. RS218 also demonstrated transcytosis of both T84 and Caco-2 cells. Four clinical isolates of E. coli K1 demonstrated levels of transcytosis of T84 cells similar to RS218. Caco-2 invasiveness correlated with length of time in tissue culture with maximum invasiveness demonstrated at 11 d in culture, when cells were polarized and differentiated.

Identification and detection of Stenotrophomonas maltophilia by rRNA-directed PCR

Stenotrophomonas maltophilia has recently emerged as an important nosocomial pathogen in immunocompromised patients, in transplant recipients, and in persons with cystic fibrosis (CF). While this organism is nonpathogenic in healthy individuals, it is increasingly associated with morbidity and mortality in susceptible populations. Recent studies have indicated that for approximately 10% of CF patients with moderate lung disease, S. maltophilia can be cultured from respiratory tract secretions. Identification of S. maltophilia can be problematic, and analysis of isolates from the Burkholderia cepacia Research Laboratory and Repository showed that several isolates presumptively identified as B. cepacia by clinical microbiology laboratories were in fact S. maltophilia. To overcome the problems associated with definitive identification, we developed species-specific PCR (SS-PCR) primers, designated SM1 and SM4, directed to the 23S rRNA gene, and tested their utility to accurately identify S. maltophilia directly from sputum. The SS-PCR was developed and tested against a panel of 112 S. maltophilia isolates collected from diverse geographic locations. To test for specificity, 43 isolates from 17 different species were analyzed. PCR with the SM1-SM4 primer pair and isolated genomic DNA as a template resulted in amplification of a band from all S. maltophilia isolates and was uniformly negative for all other species tested, yielding a sensitivity and a specificity of 100% for the SS-PCR. The utility of the SS-PCR to directly identify S. maltophilia in sputum was examined. Thirteen expectorated sputum samples from CF patients were analyzed by SS-PCR. Three samples were PCR positive, in complete concordance with the conventional laboratory culture. Thus, we have developed an SS-PCR protocol that can rapidly and accurately identify S. maltophilia isolates and which can be used for the direct detection of this organism in CF patient sputum.

Factor analysis of the Condom Use Self-Efficacy Scale among multicultural college students

The Condom Use Self-Efficacy Scale (CUSES) was administered to 447 multicultural college students. The sample consisted of 63.5% Hispanic/Latino, 17.1% African-American, 13.7% Caucasian, 4.1% other and 1.6% Asian students. The obtained scores were subjected to a principal components factor analysis with a Varimax rotation. An item designation criteria was used and three distinct factors were extracted: (1) 'Appropriation', (2) 'Sexually Transmitted Diseases' and (3) 'Partners' Disapproval'. Comparisons to the only other published factor analysis of the CUSES are made. Implications for future research using the CUSES to design AIDS education curricula for multicultural college students are discussed.

Comparison of agar diffusion methodologies for antimicrobial susceptibility testing of Pseudomonas aeruginosa isolates from cystic fibrosis patients

Pseudomonas aeruginosa is the most common pathogen infecting the lungs of patients with cystic fibrosis (CF). Improved antimicrobial chemotherapy has significantly increased the life expectancy of these patients. However, accurate susceptibility testing of P. aeruginosa isolates from CF sputum may be difficult because the organisms are often mucoid and slow growing. This study of 597 CF isolates of P. aeruginosa examined the correlation of disk diffusion and Etest (AB BIODISK, Solna, Sweden) results with a reference broth microdilution method. The rates of interpretive errors for 12 commonly used antipseudomonal antimicrobials were determined. The disk diffusion method correlated well (zone diameter versus MIC) for all of the agents tested. However, for mucoid isolates, correlation coefficients (r values) for piperacillin, piperacillin-tazobactam, and meropenem were <0.80. The Etest correlation with reference broth microdilution results (MIC versus MIC) was acceptable for all of the agents tested, for both mucoid and nonmucoid isolates. Category interpretation errors were similar for the disk diffusion and Etest methods with 0.4 and 0.1%, respectively, very major errors (false susceptibility) and 1.1 and 2.2% major errors (false resistance). Overall, both agar diffusion methods appear to be broadly acceptable for routine clinical use in susceptibility testing of CF isolates of P. aeruginosa.

Aminoglycoside-resistance mechanisms for cystic fibrosis Pseudomonas aeruginosa isolates are unchanged by long-term, intermittent, inhaled tobramycin treatment

Aminoglycoside-resistance mechanisms were characterized in Pseudomonas aeruginosa isolates from cystic fibrosis (CF) patients during a recent clinical trial of inhaled tobramycin. Impermeability, in which bacteria have reduced susceptibility to all aminoglycosides, was the predominant mode of resistance in isolates obtained both before and after 6 months of cyclic treatment with tobramycin or placebo administered by aerosol. Enzymatic resistance mechanisms were found in fewer than 10% of resistant isolates. P. aeruginosa from individual patients could be grouped on the basis of genetic relatedness. When enzymatic resistance was involved, all isolates in a group had elevated tobramycin MICs. When impermeability occurred, MICs of a genotypic group varied from susceptible to resistant. These findings suggest that impermeability resistance occurs in only a fraction of the P. aeruginosa population in lungs of persons with CF and that this form of resistance arises by a process involving multiple small changes in MIC.

Activities of tobramycin and six other antibiotics against Pseudomonas aeruginosa isolates from patients with cystic fibrosis

The in vitro activity of tobramycin was compared with those of six other antimicrobial agents against 1,240 Pseudomonas aeruginosa isolates collected from 508 patients with cystic fibrosis during pretreatment visits as part of the phase III clinical trials of tobramycin solution for inhalation. The tobramycin MIC at which 50% of isolates are inhibited (MIC(50)) and MIC(90) were 1 and 8 microg/ml, respectively. Tobramycin was the most active drug tested and also showed good activity against isolates resistant to multiple antibiotics. The isolates were less frequently resistant to tobramycin (5.4%) than to ceftazidime (11.1%), aztreonam (11.9%), amikacin (13.1%), ticarcillin (16.7%), gentamicin (19.3%), or ciprofloxacin (20.7%). For all antibiotics tested, nonmucoid isolates were more resistant than mucoid isolates. Of 56 isolates for which the tobramycin MIC was > or = 16 microg/ml and that were investigated for resistance mechanisms, only 7 (12.5%) were shown to possess known aminoglycoside-modifying enzymes; the remaining were presumably resistant by an incompletely understood mechanism often referred to as "impermeability."

Specific lipopolysaccharide found in cystic fibrosis airway Pseudomonas aeruginosa

Cystic fibrosis (CF) patients develop chronic airway infections with Pseudomonas aeruginosa (PA). Pseudomonas aeruginosa synthesized lipopolysaccharide (LPS) with a variety of penta- and hexa-acylated lipid A structures under different environmental conditions. CF patient PA synthesized LPS with specific lipid A structures indicating unique recognition of the CF airway environment. CF-specific lipid A forms containing palmitate and aminoarabinose were associated with resistance to cationic antimicrobial peptides and increased inflammatory responses, indicating that they are likely to be involved in airway disease.

Fifteen years of experience with bacterial meningitis

BACKGROUND

Introduction of Haemophilus influenzae type b (Hib) vaccines has dramatically altered the epidemiology of bacterial meningitis in children. The goal of this study was to describe these changes in a pediatric teaching hospital.

METHODS

Patient charts at Children's Hospital and Regional Medical Center, Seattle, were identified by diagnosis codes and reviewed retrospectively. The 1981 to 1995 time period was chosen to incorporate three distinct 5-year periods: before the use of unconjugated Hib vaccine; between the unconjugated and conjugate vaccines; and after the conjugate vaccines were available for routine immunization of infants.

RESULTS

Bacterial meningitis was identified in 806 cases. In 13 premature infant cases Escherichia coli was most frequently isolated (6 cases). Group B Streptococcus, E. coli and Listeria monocytogenes were the most common pathogens in 87 neonatal cases. The most common pathogens in 706 cases of childhood meningitis were H. influenzae, Streptococcus pneumoniae and Neisseria meningitidis. H. influenzae was the most common pathogen in the first two time periods (73 and 69% of childhood cases, respectively), but not so in the third period (16%).

CONCLUSIONS

A changing pattern in childhood meningitis was observed during the study period. H. influenzae cases dramatically declined, altering the relative proportions of other pathogens, S. pneumoniae and N. meningitidis. However, the number of cases caused by these latter pathogens remained steady.

Evaluation of reference dilution test methods for antimicrobial susceptibility testing of Pseudomonas aeruginosa strains isolated from patients with cystic fibrosis

The development of multidrug-resistant Pseudomonas aeruginosa in patients with cystic fibrosis (CF) is most likely a consequence of increasing life expectancy and more prolonged exposure to antibiotics. The optimal method for antibiotic susceptibility testing of CF strains, particularly mucoid P. aeruginosa strains, is unknown. Antimicrobial susceptibilities of 48 CF strains (25 mucoid) and 50 non-CF strains to 12 anti-Pseudomonas agents were tested by both agar dilution and commercially custom-prepared broth microdilution plates (PML Microbiologicals, Portland, Oreg.) in three laboratories simultaneously to determine if broth microdilution could substitute for agar dilution as the reference method in subsequent studies. Comparison of MICs generated by agar dilution and broth microdilution demonstrated correlation coefficients (r) exceeding 0.85 for all agents tested; correlation was excellent for aminoglycosides (r >/= 0.92) and very good for beta-lactam agents including agents paired with a beta-lactamase inhibitor (r >/= 0.87) and for ciprofloxacin (r = 0.86). Correlation was not improved by 48-h readings, but correlation between 24- and 48-h readings ranged between 0.91 and 0.98 for both methods. Interlaboratory variations were minimal, as the percentage of acceptable variations was 94% for both methods, and serious discords were infrequent (<2% of comparisons). However, CF strains were more likely to have serious discords than were non-CF strains (P < 0. 0001), although mucoid strains were not more likely to have serious discords than were nonmucoid strains. In this study, MICs determined by custom-prepared broth microdilution compared favorably with MICs determined by agar dilution. Thus, this broth microdilution assay can serve as a reference method and facilitate future studies to determine the optimal method for antibiotic susceptibility testing of CF strains.

Calcineurin is required for skeletal muscle hypertrophy

Molecular signaling pathways linking increases in skeletal muscle usage to alterations in muscle size have not been identified. In the present study, we tested the hypothesis that calcineurin, a calcium-regulated phosphatase recently implicated in the signaling of some forms of cardiomyopathic growth, is required to induce skeletal muscle hypertrophy and muscle fiber type conversions associated with functional overload in vivo. Administration of the specific calcineurin inhibitors cyclosporin (CsA) or FK506 to mice, for which the fast plantaris muscle was overloaded for 1-4 weeks, prevented the rapid doubling of mass and individual fiber size and the 4-20-fold increase in the number of slow fibers that characterize this condition. CsA treatment influenced the expression of muscle myofibrillar protein genes in a way reflective of fiber phenotype transformations but only in the long term of the overload condition, suggesting that the control of this growth response by calcineurin is not limited to the transcriptional activation of these muscle-specific genes. Clinically, these results provide insight to the post-surgical muscle wasting and weakness observed in recovering transplant recipients administered therapeutic dosages of these immunosuppressants.

Effect of chronic intermittent administration of inhaled tobramycin on respiratory microbial flora in patients with cystic fibrosis

Pseudomonas aeruginosa endobronchial infection causes significant morbidity and mortality among cystic fibrosis patients. Microbiology results from two multicenter, double-blind, placebo-controlled trials of inhaled tobramycin in cystic fibrosis were monitored for longitudinal changes in sputum microbial flora, antibiotic susceptibility, and selection of P. aeruginosa isolates with decreased tobramycin susceptibility. Clinical response was examined to determine whether current susceptibility standards are applicable to aerosolized administration. Treatment with inhaled tobramycin did not increase isolation of Burkholderia cepacia, Stenotrophomonas maltophilia, or Alcaligenes xylosoxidans; however, isolation of Candida albicans and Aspergillus species did increase. Although P. aeruginosa tobramycin susceptibility decreased in the tobramycin group compared with that in the placebo group, there was no evidence of selection for the most resistant isolates to become most prevalent. The definition of resistance for parenteral administration does not apply to inhaled tobramycin: too few patients had P. aeruginosa with a tobramycin MIC >/=16 microgram/mL to define a new break point on the basis of clinical response.

Comparison of isolation media for recovery of Burkholderia cepacia complex from respiratory secretions of patients with cystic fibrosis

Burkholderia cepacia selective agar (BCSA) has previously been devised for isolation of B. cepacia from respiratory secretions of patients with cystic fibrosis and tested under research laboratory conditions. Here we describe a study in which BCSA, oxidation-fermentation polymyxin bacitracin lactose agar (OFPBL), and Pseudomonas cepacia agar (PCA) were compared in routine culture procedures for the ability to grow B. cepacia and inhibit other organisms. Three hundred twenty-eight specimens from 209 patients at two pediatric centers and 328 specimens from 109 adults were tested. Plates were inoculated, incubated, and read for quality and quantity of growth at 24, 48, and 72 h. Five (1.5%) specimens from 4 (1.9%) children and 75 (22.9%) specimens from 16 (14.7%) adults grew B. cepacia complex. At 24, 48, and 72 h, BCSA achieved 43, 93, and 100% detection, respectively; OFPBL achieved 26, 84, and 96%, respectively; and PCA achieved 33, 74, and 84% detection, respectively. Quality was assessed as pinpoint or good growth. At 24 h, most cultures growing B. cepacia complex had pinpoint colonies. By 48 and 72 h, 48 and 69% of B. cepacia complex cultures, respectively, had good growth on BCSA, while on OFPBL 19 and 30%, respectively, had good growth and on PCA 11 and 18%, respectively, had good growth. BCSA was superior to OFPBL and PCA in suppressing organisms other than B. cepacia complex; 40 non-B. cepacia complex organisms were isolated from BCSA, 263 were isolated from OFPBL, and 116 were isolated from PCA. We conclude that BCSA is superior to OFPBL and PCA in its ability to support the growth of B. cepacia complex and to suppress other respiratory organisms.

Microbiology of sputum from patients at cystic fibrosis centers in the United States

During a phase III national collaborative study of aerosolized tobramycin (1 July 1995 through 30 September 1996), the microbiology of specimens from 595 patients at 69 cystic fibrosis (CF) centers was examined. Samples from three screening visits were processed in a single laboratory by means of standardized techniques for identification and susceptibility testing. From 1,753 pretreatment specimens, 5,128 pathogens were isolated (average, 2.9/specimen). Of the 3,936 Pseudomonas aeruginosa isolates, 56.7% were mucoid. The specimens of 125 patients (21.0%) yielded tobramycin-resistant P. aeruginosa (213 isolates); 61 (10.3%), Stenotrophomonas maltophilia; and 52 (8.7%), Alcaligenes xylosoxidans. Isolation of Burkholderia cepacia was an exclusion criterion. Only visit 3 sputum samples were cultured for gram-positive organisms and fungi (n = 465 patients); samples from 201 patients (43.2%) yielded Staphylococcus aureus (18.8% of isolates were oxacillin-resistant), and those from 114 (24.5%) yielded an Aspergillus species. Compared with the Cystic Fibrosis Foundation Patient Registry, the current study identified many more patients colonized with S. maltophilia, A. xylosoxidans, Aspergillus species, and oxacillin-resistant S. aureus, suggesting the utility of standardized processing of CF specimens.