Risk factors for severe respiratory syncytial virus-associated respiratory tract infection in a high HIV prevalence setting, South Africa, 2012 - 2018.

BACKGROUND: Identifying risk factors for respiratory syncytial virus (RSV)-associated severe acute respiratory illness (SARI) will assist with targeting vaccine interventions. METHODS: Using surveillance data from South Africa (2012-2018), we compared the characteristics of individuals with RSV-associated influenza-like illness (ILI) (reference group) to those with RSV-associated SARI to describe factors associated with SARI using a multivariable analysis. RESULTS: RSV was detected in 6% (483/7792) of ILI cases and 15% (844/5672) of SARI cases. Factors associated with SARI in children included age < 2 months, compared to age 2-4 years (adjusted odds ratio (aOR) 54.4; 95% confidence interval (CI) 23.5-125.8), malnutrition (aOR 1.9; 95% CI 1.2-3.2), prematurity (aOR 2.4; 95% CI 1.3-4.6) and living with HIV (LWH) (aOR 22.5; 95% CI 2.9-174.3). In individuals ≥ 5 years, factors associated with SARI included age ≥ 65 years compared to age 5-24 years (aOR 10.7; 95% CI 1.1-107.5), symptom duration ≥ 5 days (aOR 2.7; 95% CI 1.1-6.3), underlying illness (aOR 2.7; 95% CI 1.5-26.1) and LWH (aOR 16.8, 95% CI: 4.8-58.2). CONCLUSION: Individuals at the extremes of age and those with identified risk factors might benefit most from RSV prevention interventions. CLINICAL TRIAL NUMBER: Not applicable, this is not a clinical trial.

Unmasking Pneumococcal Carriage in a High Human Immunodeficiency Virus (HIV) Prevalence Population in two Community Cohorts in South Africa, 2016-2018: The PHIRST Study.

BACKGROUND: Longitudinal pneumococcus colonization data in high human immunodeficiency virus (HIV) prevalence settings following pneumococcal conjugate vaccine introduction are limited. METHODS: In 327 randomly selected households, 1684 individuals were enrolled and followed-up for 6 to 10 months during 2016 through 2018 from 2 communities. Nasopharyngeal swabs were collected twice weekly and tested for pneumococcus using quantitative lytA real-time polymerase chain reaction. A Markov model was fitted to the data to define the start and end of an episode of colonization. We assessed factors associated with colonization using logistic regression. RESULTS: During the study period, 98% (1655/1684) of participants were colonized with pneumococcus at least once. Younger age (<5 years: adjusted odds ratio [aOR], 14.1; 95% confidence [CI], 1.8-111.3, and 5-24 years: aOR, 4.8, 95% CI, 1.9-11.9, compared with 25-44 years) and HIV infection (aOR, 10.1; 95% CI, 1.3-77.1) were associated with increased odds of colonization. Children aged <5 years had fewer colonization episodes (median, 9) than individuals ≥5 years (median, 18; P < .001) but had a longer episode duration (<5 years: 35.5 days; interquartile range, 17-88) vs. ≥5 years: 5.5 days (4-12). High pneumococcal loads were associated with age (<1 year: aOR 25.4; 95% CI, 7.4-87.6; 1-4 years: aOR 13.5, 95% CI 8.3-22.9; 5-14 years: aOR 3.1, 95% CI, 2.1-4.4 vs. 45-65 year old patients) and HIV infection (aOR 1.7; 95% CI 1.2-2.4). CONCLUSIONS: We observed high levels of pneumococcus colonization across all age groups. Children and people with HIV were more likely to be colonized and had higher pneumococcal loads. Carriage duration decreased with age highlighting that children remain important in pneumococcal transmission.

Transient increased risk of influenza infection following RSV infection in South Africa: findings from the PHIRST study, South Africa, 2016-2018.

BACKGROUND: Large-scale prevention of respiratory syncytial virus (RSV) infection may have ecological consequences for co-circulating pathogens, including influenza. We assessed if and for how long RSV infection alters the risk for subsequent influenza infection. METHODS: We analysed a prospective longitudinal cohort study conducted in South Africa between 2016 and 2018. For participating households, nasopharyngeal samples were taken twice weekly, irrespective of symptoms, across three respiratory virus seasons, and real-time polymerase chain reaction (PCR) was used to identify infection with RSV and/or influenza. We fitted an individual-level hidden Markov transmission model in order to estimate RSV and influenza infection rates and their interdependence. RESULTS: Of a total of 122,113 samples collected, 1265 (1.0%) were positive for influenza and 1002 (0.8%) positive for RSV, with 15 (0.01%) samples from 12 individuals positive for both influenza and RSV. We observed a 2.25-fold higher incidence of co-infection than expected if assuming infections were unrelated. We estimated that infection with influenza is 2.13 (95% CI 0.97-4.69) times more likely when already infected with, and for a week following, RSV infection, adjusted for age. This equates to 1.4% of influenza infections that may be attributable to RSV in this population. Due to the local seasonality (RSV season precedes the influenza season), we were unable to estimate changes in RSV infection risk following influenza infection. CONCLUSIONS: We find no evidence to suggest that RSV was associated with a subsequent reduced risk of influenza infection. Instead, we observed an increased risk for influenza infection for a short period after infection. However, the impact on population-level transmission dynamics of this individual-level synergistic effect was not measurable in this setting.

Epidemiology of Pertussis in Individuals of All Ages Hospitalized With Respiratory Illness in South Africa, January 2013-December 2018.

BACKGROUND: Policy recommendations on pertussis vaccination need to be guided by data, which are limited from low- and middle-income countries. We aimed to describe the epidemiology of pertussis in South Africa, a country with high human immunodeficiency virus (HIV) prevalence and routine pertussis vaccination for 6 decades including the acellular vaccine since 2009. METHODS: Hospitalized patients of all ages were enrolled at 5 sentinel sites as part of a pneumonia surveillance program from January 2013 through December 2018. Nasopharyngeal specimens and induced sputum were tested by polymerase chain reaction (PCR) for Bordetella pertussis. In addition, demographic and clinical information were collected. Incidence rates were calculated for 2013-2016, and multivariable logistic regression performed to identify factors associated with pertussis. RESULTS: Over the 6-year period 19 429 individuals were enrolled, of which 239 (1.2%) tested positive for B. pertussis. Detection rate was highest in infants aged <6 months (2.8%, 155/5524). Mean annual incidence was 17 cases per 100 000 population, with the highest incidence in children <1 year of age (228 per 100 000). Age-adjusted incidence was 65.9 per 100 000 in HIV-infected individuals compared to 8.5 per 100 000 in HIV-uninfected individuals (risk ratio 30.4, 95% confidence interval: 23.0-40.2). Ten individuals (4.2%) with pertussis died; of which 7 were infants aged <6 months and 3 were immunocompromised adults. CONCLUSIONS: Pertussis continues to be a significant cause of illness and hospitalization in South Africa, despite routine vaccination. The highest burden of disease and death occurred in infants; however, HIV-infected adults were also identified as an important group at risk of B. pertussis infection.

Rift Valley Fever Virus Seroprevalence among Humans, Northern KwaZulu-Natal Province, South Africa, 2018-2019.

We detected Rift Valley fever virus (RVFV) IgM and IgG in human serum samples collected during 2018-2019 in northern KwaZulu-Natal Province, South Africa. Our results show recent RVFV circulation and likely RVFV endemicity in this tropical coastal plain region of South Africa in the absence of apparent clinical disease.

Proposal for Human Respiratory Syncytial Virus Nomenclature below the Species Level.

Human respiratory syncytial virus (HRSV) is the leading viral cause of serious pediatric respiratory disease, and lifelong reinfections are common. Its 2 major subgroups, A and B, exhibit some antigenic variability, enabling HRSV to circulate annually. Globally, research has increased the number of HRSV genomic sequences available. To ensure accurate molecular epidemiology analyses, we propose a uniform nomenclature for HRSV-positive samples and isolates, and HRSV sequences, namely: HRSV/subgroup identifier/geographic identifier/unique sequence identifier/year of sampling. We also propose a template for submitting associated metadata. Universal nomenclature would help researchers retrieve and analyze sequence data to better understand the evolution of this virus.

Decline of influenza and respiratory syncytial virus detection in facility-based surveillance during the COVID-19 pandemic, South Africa, January to October 2020.

BackgroundIn South Africa, COVID-19 control measures to prevent SARS-CoV-2 spread were initiated on 16 March 2020. Such measures may also impact the spread of other pathogens, including influenza virus and respiratory syncytial virus (RSV) with implications for future annual epidemics and expectations for the subsequent northern hemisphere winter.MethodsWe assessed the detection of influenza and RSV through facility-based syndromic surveillance of adults and children with mild or severe respiratory illness in South Africa from January to October 2020, and compared this with surveillance data from 2013 to 2019.ResultsFacility-based surveillance revealed a decline in influenza virus detection during the regular season compared with previous years. This was observed throughout the implementation of COVID-19 control measures. RSV detection decreased soon after the most stringent COVID-19 control measures commenced; however, an increase in RSV detection was observed after the typical season, following the re-opening of schools and the easing of measures.ConclusionCOVID-19 non-pharmaceutical interventions led to reduced circulation of influenza and RSV in South Africa. This has limited the country's ability to provide influenza virus strains for the selection of the annual influenza vaccine. Delayed increases in RSV case numbers may reflect the easing of COVID-19 control measures. An increase in influenza virus detection was not observed, suggesting that the measures may have impacted the two pathogens differently. The impact that lowered and/or delayed influenza and RSV circulation in 2020 will have on the intensity and severity of subsequent annual epidemics is unknown and warrants close monitoring.

The Fraction of Rhinovirus Detections Attributable to Mild and Severe Respiratory Illness in a Setting of High Human Immunodeficiency Virus Prevalence, South Africa, 2013-2015.

BACKGROUND: The association of rhinovirus (RV) detection to illness is poorly understood. METHODS: We enrolled case patients hospitalized with severe respiratory illness (SRI) at 2 hospitals and outpatients with influenza-like illness (ILI) and asymptomatic individuals (controls) from 2 affiliated clinics during 2013-2015. We compared the RV prevalence among ILI and SRI cases to those of controls stratified by human immunodeficiency virus (HIV) serostatus using penalized logistic regression. The attributable fraction (AF) was calculated. RESULTS: During 2013-2015, RV was detected in 17.4% (368/2120), 26.8% (979/3654), and 23.0% (1003/4360) of controls, ILI cases, and SRI cases, respectively. The RV AF (95% confidence interval) was statistically significant among children aged <5 years (ILI: 44.6% [30.7%-55.7%] and SRI: 50.3% [38.6%-59.9%]; P < .001) and individuals aged ≥5 years (ILI: 62.9% [54.4%-69.8%] and SRI: 51.3% [38.7%-61.3%]; P < .001) as well as among HIV-infected (ILI: 59.9% [45.8%-70.3%] and SRI: 39.8% [22.3%-53.3%]; P < .001) and HIV-uninfected (ILI: 53.6% [44.7%-61.1%] and SRI: 55.3% [45.6%-63.2%]; P < .001) individuals. CONCLUSIONS: Although RV detection was common among controls, it was also associated with a substantial proportion of clinical illness across age groups, irrespective of HIV status.

Household Transmission of Seasonal Influenza From HIV-Infected and HIV-Uninfected Individuals in South Africa, 2013-2014.

BACKGROUND: We estimated the household secondary infection risk (SIR) and serial interval (SI) for influenza transmission from HIV-infected and HIV-uninfected index cases. METHODS: Index cases were the first symptomatic person in a household with influenza-like illness, testing influenza positive on real-time reverse transcription polymerase chain reaction (rRT-PCR). Nasopharyngeal swabs collected from household contacts every 4 days were tested by rRT-PCR. Factors associated with SIR were evaluated using logistic regression. RESULTS: We enrolled 28 HIV-infected and 57 HIV-uninfected index cases. On multivariable analysis, HIV-infected index cases were less likely to transmit influenza to household contacts (odds ratio [OR] 0.2; 95% confidence interval [CI], 0.1-0.6; SIR 16%, 18/113 vs 27%, 59/220). Factors associated with increased SIR included index age group 1-4 years (OR 3.6; 95% CI, 1.2-11.3) and 25-44 years (OR 8.0; 95% CI, 1.8-36.7), and contact age group 1-4 years (OR 3.5; 95% CI, 1.2-10.3) compared to 5-14 years, and sleeping with index case (OR 2.7; 95% CI, 1.3-5.5). HIV infection of index case was not associated with SI. CONCLUSIONS: HIV-infection was not associated with SI. Increased infectiousness of HIV-infected individuals is likely not an important driver of community influenza transmission.

The Role of Human Immunodeficiency Virus in Influenza- and Respiratory Syncytial Virus-associated Hospitalizations in South African Children, 2011-2016.

BACKGROUND: Data describing influenza- or respiratory syncytial virus (RSV)-associated hospitalized illness in children aged <5 years in Africa are limited. METHODS: During 2011-2016, we conducted surveillance for severe respiratory illness (SRI) in children aged <5 years in 3 South African hospitals. Nasopharyngeal aspirates were tested for influenza and RSV using real-time reverse transcription polymerase chain reaction. We estimated rates of influenza- and RSV-associated hospitalized SRI by human immunodeficiency virus (HIV) status and compared children who tested positive for influenza vs RSV using multivariable penalized logistic regression. RESULTS: Among 3650 hospitalized children, 203 (5.6%) tested positive for influenza viruses, 874 (23.9%) for RSV, and 19 (0.5%) for both. The median age of children hospitalized with influenza was 13.9 months vs 4.4 months for RSV (P < .01). Annual influenza-associated hospitalization rates per 100000 were highest among infants aged 6-11 months (545; 95% confidence interval [CI], 409-703), while RSV-associated hospitalization rates were highest in infants aged 0-2 months (6593; 95% CI, 5947-7217). HIV exposure was associated with increased incidence of influenza- and RSV-associated hospitalization in infants aged 0-5 months, with relative risk (RR) 2.2 (95% CI, 1.4-3.4) and 1.4 (95% CI, 1.3-1.6), respectively. HIV infection was associated with increased incidence of influenza- and RSV-associated hospitalization in all age groups; RR 2.7 (95% CI, 2.0-3.5) and 3.8 (95% CI, 3.1-4.8), respectively. CONCLUSIONS: Influenza- and RSV-associated hospitalizations are common among South African infants. HIV infection and HIV exposure in infants increase risk of influenza- and RSV-associated hospitalization.

The Impact of Human Immunodeficiency Virus Exposure on Respiratory Syncytial Virus-associated Severe Respiratory Illness in South African Infants, 2011-2016.

From 2011 through 2016, we conducted surveillance for severe respiratory illness in infants. Human immunodeficiency virus exposure significantly increased the risk of respiratory syncytial virus (RSV)-associated hospitalization in infants aged <5 months. More than 60% of RSV-associated hospitalizations occurred in the first 4 months of life and may be preventable through maternal vaccination or birth-dose monoclonal antibody.

Quantifying How Different Clinical Presentations, Levels of Severity, and Healthcare Attendance Shape the Burden of Influenza-associated Illness: A Modeling Study From South Africa.

BACKGROUND: Burden estimates of medically and nonmedically attended influenza-associated illness across syndromes and levels of severity are lacking. METHODS: We estimated the national burden of medically and nonmedically attended influenza-associated illness among individuals with different clinical presentations (all-respiratory, all-circulatory, and nonrespiratory/noncirculatory) and levels of severity (mild, fatal, and severe, nonfatal) using a combination of case-based (from laboratory-confirmed influenza surveillance) and ecological studies, as well as data from healthcare utilization surveys in South Africa during 2013-2015. In addition, we compared estimates of medically attended influenza-associated respiratory illness, obtained from case-based and ecological studies. Rates were reported per 100 000 individuals in the population. RESULTS: The estimated mean annual number of influenza-associated illness episodes was 10 737 847 (19.8% of 54 096 705 inhabitants). Of these episodes, 10 598 138 (98.7%) were mild, 128 173 (1.2%) were severe, nonfatal, and 11 536 (0.1%) were fatal. There were 2 718 140 (25.6%) mild, 56 226 (43.9%) severe, nonfatal, and 4945 (42.8%) medically attended should be after fatal episodes. Influenza-associated respiratory illness accounted for 99.2% (10 576 146) of any mild, 65.5% (83 941) of any severe, nonfatal, and 33.7% (3893) of any fatal illnesses. Ecological and case-based estimates of medically attended, influenza-associated, respiratory mild (rates: ecological, 1778.8, vs case-based, 1703.3; difference, 4.4%), severe, nonfatal (rates: ecological, 88.6, vs case-based, 75.3; difference, 15.0%), and fatal (rates: ecological, 3.8, vs case-based, 3.5; difference, 8.4%) illnesses were similar. CONCLUSIONS: There was a substantial burden of influenza-associated symptomatic illness, including severe, nonfatal and fatal illnesses, and a large proportion was nonmedically attended. Estimates, including only influenza-associated respiratory illness, substantially underestimated influenza-associated, severe, nonfatal and fatal illnesses. Ecological and case-based estimates were found to be similar for the compared categories.

Heterogeneity in influenza seasonality and vaccine effectiveness in Australia, Chile, New Zealand and South Africa: early estimates of the 2019 influenza season.

We compared 2019 influenza seasonality and vaccine effectiveness (VE) in four southern hemisphere countries: Australia, Chile, New Zealand and South Africa. Influenza seasons differed in timing, duration, intensity and predominant circulating viruses. VE estimates were also heterogeneous, with all-ages point estimates ranging from 7-70% (I2: 33%) for A(H1N1)pdm09, 4-57% (I2: 49%) for A(H3N2) and 29-66% (I2: 0%) for B. Caution should be applied when attempting to use southern hemisphere data to predict the northern hemisphere influenza season.

Influenza Viral Shedding in a Prospective Cohort of HIV-Infected and Uninfected Children and Adults in 2 Provinces of South Africa, 2012-2014.

Background: Prolonged shedding of influenza viruses may be associated with increased transmissibility and resistance mutation acquisition due to therapy. We compared duration and magnitude of influenza shedding between human immunodeficiency virus (HIV)-infected and -uninfected individuals. Methods: A prospective cohort study during 3 influenza seasons enrolled patients with influenza-like illness and a positive influenza rapid test. Influenza viruses were detected by real-time reverse transcription polymerase chain reaction. Weibull accelerated failure time regression models were used to describe influenza virus shedding. Mann-Whitney U tests explored initial influenza viral loads (VL). Results: Influenza virus shedding duration was similar in 65 HIV-infected (6 days; interquartile range [IQR] 3-10) and 176 HIV-uninfected individuals (7 days; IQR 4-11; P = .97), as was initial influenza VL (HIV-uninfected 5.28 ± 1.33 log10 copies/mL, HIV-infected 4.73 ± 1.68 log10 copies/mL; P = .08). Adjusted for age, HIV-infected individuals with low CD4 counts shed influenza virus for longer than those with higher counts (adjusted hazard ratio 3.55; 95% confidence interval, 1.05-12.08). Discussion: A longer duration of influenza virus shedding in HIV-infected individuals with low CD4 counts may suggest a possible increased risk for transmission or viral evolution in severely immunocompromised individuals. HIV-infected individuals should be prioritized for annual influenza immunization.

Epidemiology and Molecular Identification and Characterization of Mycoplasma pneumoniae, South Africa, 2012-2015.

During 2012-2015, we tested respiratory specimens from patients with severe respiratory illness (SRI), patients with influenza-like illness (ILI), and controls in South Africa by real-time PCR for Mycoplasma pneumoniae, followed by culture and molecular characterization of positive samples. M. pneumoniae prevalence was 1.6% among SRI patients, 0.7% among ILI patients, and 0.2% among controls (p<0.001). Age <5 years (adjusted odd ratio 7.1; 95% CI 1.7-28.7) and HIV infection (adjusted odds ratio 23.8; 95% CI 4.1-138.2) among M. pneumonia-positive persons were associated with severe disease. The detection rate attributable to illness was 93.9% (95% CI 74.4%-98.5%) in SRI patients and 80.7% (95% CI 16.7%-95.6%) in ILI patients. The hospitalization rate was 28 cases/100,000 population. We observed the macrolide-susceptible M. pneumoniae genotype in all cases and found P1 types 1, 2, and a type 2 variant with multilocus variable number tandem repeat types 3/6/6/2, 3/5/6/2, and 4/5/7/2.

Severity of Respiratory Syncytial Virus Lower Respiratory Tract Infection With Viral Coinfection in HIV-Uninfected Children.

Background: Molecular diagnostics enable sensitive detection of respiratory viruses, but their clinical significance remains unclear in pediatric lower respiratory tract infection (LRTI). We aimed to determine whether viral coinfections increased life-threatening disease in a large cohort. Methods: Molecular testing was performed for respiratory viruses in nasopharyngeal aspirates collected from children aged <5 years within 24 hours of hospital admission during sentinel surveillance for severe acute respiratory illness (SARI) hospitalization conducted in South Africa during February 2009-December 2013. The primary outcome was life-threatening disease, defined as mechanical ventilation, intensive care unit admission, or death. Results: Of 2322 HIV-uninfected children with respiratory syncytial virus (RSV)-associated LRTI, 1330 (57.3%) had RSV monoinfection, 38 (1.6%) had life-threatening disease, 575 (24.8%) had rhinovirus, 347 (14.9%) had adenovirus (ADV), and 30 (1.3%) had influenza virus. RSV and any other viral coinfection was not associated with severe disease, ADV coinfection had increased odds of life-threatening disease (adjusted OR, 3.4; 95% CI, 1.6-7.2; P = .001), and influenza coinfection had increased odds of life-threatening disease and prolonged length of stay (adjusted OR, 2.1; 95% CI, 1.0-4.5; P = .05) compared with RSV monoinfection. Conclusions: RSV coinfection with any respiratory virus is not associated with more severe disease when compared to RSV alone in this study. However, increased life-threatening disease in RSV-ADV and RSV-influenza coinfection warrants further study.

Attributable Fraction of Influenza Virus Detection to Mild and Severe Respiratory Illnesses in HIV-Infected and HIV-Uninfected Patients, South Africa, 2012-2016.

The attributable fraction (AF) of influenza virus detection to illness has not been described for patients in different age groups or with different HIV infection statuses. We compared the age group-specific prevalence of influenza virus infection among patients with influenza-like illness (ILI) or severe acute or chronic respiratory illness (SARI and SCRI, respectively) with that among controls, stratified by HIV serostatus. The overall AF for influenza virus detection to illness was 92.6% for ILI, 87.4% for SARI, and 86.2% for SCRI. Among HIV-uninfected patients, the AF for all syndromes was highest among persons <1 and >65 years of age and lowest among persons 25-44 years of age; this trend was not observed among HIV-infected patients. Overall, influenza viruses when detected in patients with ILI, SARI, or SCRI are likely attributable to illness. This finding is particularly likely among children and the elderly irrespective of HIV serostatus and among HIV-infected persons irrespective of age.

Enterovirus genotypes among patients with severe acute respiratory illness, influenza-like illness, and asymptomatic individuals in South Africa, 2012-2014.

Enteroviruses can cause outbreaks of severe acute respiratory illness (SARI) and EV-A, -B, -C, and -D species have different pathogenic profiles and circulation patterns. We aimed to characterize and determine the prevalence of enterovirus genotypes among South African patients with respiratory illness and controls during June 2012 to July 2014. Syndromic SARI and influenza-like illness (ILI) surveillance was performed at two sentinel sites. At each site nasopharyngeal/oropharyngeal specimens were collected from SARI and ILI patients as well as controls. Specimens were tested for enterovirus by real-time PCR. Positive specimens were further genotyped by sequencing a region of the VP1 gene. The prevalence of enterovirus was 5.8% (87/1494), 3.4% (103/3079), and 3.4% (46/1367) among SARI, ILI, and controls, respectively (SARI/controls, P = 0.002 and ILI/control, P = 0.973). Among the 101/236 (42.8%) enterovirus-positive specimens that could be genotyped, we observed a high diversity of circulating enterovirus genotypes (a total of 33 genotypes) from all four human enterovirus species with high prevalence of Enterovirus-B (60.4%; 61/101) and Enterovirus-A (21.8%; 22/101) compared to Enterovirus-C (10.9%; 11/101) and Enterovirus-D (6.9%; 7/101) (P = 0.477). Of the enterovirus genotypes identified, Echovirus 30 (9.9%, 10/101), Coxsackie virus B5 (7.9%, 8/101) and Enterovirus-D68 (6.9%, 7/101) were most prevalent. There was no difference in disease severity (SARI or ILI compared to controls) between the different enterovirus species (P = 0.167). We observed a high number of enterovirus genotypes in patients with respiratory illness and in controls from South Africa with no disease association of EV species with disease severity.

Legionnaires' Disease in South Africa, 2012-2014.

During June 2012-September 2014, we tested patients with severe respiratory illness for Legionella spp. infection and conducted a retrospective epidemiologic investigation. Of 1,805 patients tested, Legionella was detected in samples of 21 (1.2%); most were adults who had HIV or tuberculosis infections and were inappropriately treated for Legionella.