More than 10 years of unrecognized nosocomial transmission of legionnaires' disease among transplant patients

OBJECTIVE: To investigate a cluster of cases of legionnaires' disease among patients at a hospital. SETTING: A university hospital that is a regional transplant center. DESIGN: Retrospective review of microbiology and serology data from the hospital laboratories and prospective surveillance via the radiology department; a case-control study and environmental sampling within the hospital and from nearby cooling towers. RESULTS: Diagnosis of seven cases of legionnaires' disease in the first 9 months of 1996 led to recognition of a nosocomial outbreak that may have begun as early as 1979. Review of charts from 1987 through September 1996 identified 25 culture-confirmed cases of nosocomial or possibly nosocomial legionnaires' disease, including 18 in bone marrow and heart transplant patients. Twelve patients (48%) died. During the first 9 months of 1996, the attack rate was 6% among cardiac and bone marrow transplant patients. For cases that occurred before 1996, intubation was associated with increased risk for disease. High-dose corticosteroid medication was strongly associated with the risk for disease, but other immunosuppressive therapy or cancer chemotherapy was not. Several species and serogroups of Legionella were isolated from numerous sites in the hospital's potable water system. Six of seven available clinical isolates were identical and were indistinguishable from environmental isolates by pulsed-field gel electrophoresis. Initial infection control measures failed to interrupt nosocomial acquisition of infection. After extensive modifications to the water system, closely monitored repeated hyperchlorinations, and reduction of patient exposures to aerosols, transmission was interrupted. No cases have been identified since September 1996. CONCLUSIONS: Legionella can colonize hospital potable water systems for long periods of time, resulting in an ongoing risk for patients, especially those who are immunocompromised. In this hospital, nosocomial transmission possibly occurred for more than 17 years and was interrupted in 1996, after a sudden increase in incidence led to its recognition. Hospitals specializing in the care of immunocompromised patients (eg, transplant centers) should prioritize surveillance for cases of legionnaires' disease. Aggressive control measures can interrupt transmission of this disease successfully.     

Outbreak of pneumonia in a long-term care facility: antecedent human parainfluenza virus 1 infection may predispose to bacterial pneumonia

OBJECTIVES: To determine the causes of an outbreak of lobar pneumonia. DESIGN: Matched (1:2) case-control study. SETTING: A 70-bed chronic care facility for older people. PARTICIPANTS: Residents of the facility. RESULTS: Ten residents developed pneumonia over a 10-day period. Two residents died. One case-patient had Streptococcus pneumoniae bacteremia; another had polymerase chain reaction evidence of S. pneumoniae infection. No other etiologic agent was identified. Only four of 10 case-patients had received routine diagnostic cultures of blood or sputum before the administration of antibiotics. Symptoms of upper respiratory illness (URI) among residents before the pneumonia outbreak corresponded with elevation of antibodies to human parainfluenza virus 1 (HPIV1). In a matched case-control study, six of 10 case-patients, compared with five of 20 controls, had symptoms of URI during the preceding month (matched odds ratio (MOR) = 4.5, 95% CI = 0.8-33). Nine case-patients had serum available, and five of these had both serologic evidence of recent HPIV1 infection and recent URI, compared with two of 18 controls (MOR = 9.0, 95% CI = 1.2-208). Only three residents had documentation of pneumococcal vaccination. CONCLUSIONS: Noninfluenza viral infections may play a role in the pathogenesis of some bacterial pneumonias. S. pneumoniae was the cause of at least two pneumonias; lack of preantibiotic cultures may have interfered with isolation of S. pneumoniae in others. Recent HPIV1 infection was epidemiologically linked to subsequently developing pneumonia. Spread of HPIV1 in the facility may have contributed to increased susceptibility to S. pneumoniae and, potentially, to other bacterial pathogens.     

Impact of technology on the emergence of infectious diseases

Not all new threats from microorganisms result from changes within the microorganisms themselves. Innovative, progressive changes introduced by humans can provide advantages to microbes which, in turn, threaten the population. In most cases, advantages are provided unwittingly as a by-product of technology, and can be corrected when the problem is recognized. An end to progress is neither a practical nor desirable solution. Alertness (i.e., surveillance) for the emergence of new diseases and to changes in disease patterns will continue to be critical to ensure timely implementation of public health measures.     

Molecular characterization of a globally distributed lineage of serotype 12F Streptococcus pneumoniae causing invasive disease

These studies have identified a major genetic lineage of capsule serotype 12F Streptococcus pneumoniae, which has maintained two different types of the pneumococcal surface protein A (PspA) virulence factor and caused invasive disease in geographically disjoint locations. Twenty outbreak strains from a Texas jail and Maryland day care center and 16 reference strains from Texas, Maryland, Washington, Michigan, Oklahoma, Missouri, Alaska, and Australia were examined. Although the Texas and Maryland outbreak strains were indistinguishable by IS1167 and boxA genotyping procedures, all strains examined were members of a genetically similar lineage. The microevolutionary history of pspA differed from that of the overall genetic background of the strains. Taken together, these findings suggested that the Texas and Maryland outbreaks were caused by different clones of a major genetic lineage of serotype 12F pneumococci, within which at least one PspA has been acquired via localized genetic recombination.     

Using Iterated Bagging to Debias Regressions

Breiman (Machine Learning, 26(2), 123–140) showed that bagging could effectively reduce the variance of regression predictors, while leaving the bias relatively unchanged. A new form of bagging we call iterated bagging is effective in reducing both bias and variance. The procedure works in stages—the first stage is bagging. Based on the outcomes of the first stage, the output values are altered; and a second stage of bagging is carried out using the altered output values. This is repeated until a simple rule stops the process. The method is tested using both trees and nearest neighbor regression methods. Accuracy on the Boston Housing data benchmark is comparable to the best of the results gotten using highly tuned and compute- intensive Support Vector Regression Machines. Some heuristic theory is given to clarify what is going on. Application to two-class classification data gives interesting results.     

Spiranic β-Ketoesters and their Pyrazolinones

A number of spiranic β-ketoesters have been prepared and transformed into the corresponding phenylpyrazolinones by condensation with phenylhydrazine. The structures of the β-ketoesters and the pyrazolinones, derived from their UV, IR and NMR spectra, are discussed.     

Bagging Predictors

Bagging predictors is a method for generating multiple versions of a predictor and using these to get an aggregated predictor. The aggregation averages over the versions when predicting a numerical outcome and does a plurality vote when predicting a class. The multiple versions are formed by making bootstrap replicates of the learning set and using these as new learning sets. Tests on real and simulated data sets using classification and regression trees and subset selection in linear regression show that bagging can give substantial gains in accuracy. The vital element is the instability of the prediction method. If perturbing the learning set can cause significant changes in the predictor constructed, then bagging can improve accuracy.     

Pasting Small Votes for Classification in Large Databases and On-Line

Many databases have grown to the point where they cannot fit into the fast memory of even large memory machines, to say nothing of current workstations. If what we want to do is to use these data bases to construct predictions of various characteristics, then since the usual methods require that all data be held in fast memory, various work-arounds have to be used. This paper studies one such class of methods which give accuracy comparable to that which could have been obtained if all data could have been held in core and which are computationally fast. The procedure takes small pieces of the data, grows a predictor on each small piece and then pastes these predictors together. A version is given that scales up to terabyte data sets. The methods are also applicable to on-line learning.     

Stacked Regressions

Stacking regressions is a method for forming linear combinations of different predictors to give improved prediction accuracy. The idea is to use cross-validation data and least squares under non negativity constraints to determine the coefficients in the combination. Its effectiveness is demonstrated in stacking regression trees of different sizes and in a simulation stacking linear subset and ridge regressions. Reasons why this method works are explored. The idea of stacking originated with Wolpert (1992).