Disseminated Infection With FREE

MMWR. 2002;51:501-502

Persons with advanced human immunodeficiency virus (HIV)-1 infection are susceptible to disseminated mycobacterial infections. In the United States, most such infections are caused by Mycobacterium avium or M. intracellulare (i.e., M. avium complex [MAC]). In less developed countries, M. tuberculosis is equally or more prevalent than MAC in persons with HIV-1 infection.¹ Other mycobacterial species have been reported to cause disseminated infection in HIV-infected persons, including Simiae-Avium (SAV) group mycobacteria. SAV group organisms share characteristics of M. avium and M. simiae.² Although disseminated (i.e., the isolation of a mycobacterial species from the blood) infection with M. simiae has been reported in HIV-infected persons,^3- 6 another distinct species within the SAV group, M. triplex, was characterized in 1996.⁷ Two cases of disseminated infection caused by M. triplex have been reported in HIV-1-positive persons.^8- 9 This report describes four HIV-infected patients from Bangkok, Thailand, and Lilongwe, Malawi, who were infected with SAV group organisms. Because different mycobacterial species are not susceptible uniformly to antimycobacterial agents, accurate identification of mycobacterial species causing an infection is crucial for directing appropriate therapy.

These infections were detected during prospective blood culture studies of febrile, adult inpatients in these two countries.^1,10 The Bangkok study was conducted at an infectious diseases hospital during February-March 1997¹⁰; the Lilongwe study was conducted at a general hospital during August-September 1997.¹ In both studies, adults (aged ≥18 years) admitted consecutively with fever (oral temperature ≥100°F [≥38°C] in Bangkok and axillary temperature ≥99°F [≥37.5°C] in Malawi) were recruited within 12 hours of hospital admission. After informed consent was obtained, patients gave a full medical history and underwent a comprehensive physical examination. Blood was drawn for HIV-1 testing and mycobacterial culture. All mycobacterial isolates were sent to Duke University Medical Center for confirmation and identification. M. tuberculosis complex and M. avium complex isolates were identified by using AccuPROBE (Gen-Probe, San Diego, California) DNA probes and biochemical tests. Isolates of uncommon Mycobacterium spp. (e.g., M. simiae) were sent to North Carolina State Public Health Laboratory and the Mycobacteria Reference Laboratory at CDC for further characterization and confirmation by high performance liquid chromatography analysis of mycolic acids. Personnel at both laboratories read all chromatograms visually. Susceptibilities of the isolates to antituberculous drugs were performed at CDC using methodology established for M. tuberculosis. Of 480 patients evaluated, four (two from Bangkok and two from Lilongwe) were found to have disseminated infection with SAV group mycobacteria, later identified as M. simiae.

Both patients had positive serology for HIV-1 antibody. Neither was receiving antiretroviral or antimycobacterial therapy. Patient 1, a man aged 32 years, presented with fever, cachexia, and diarrhea of 3 months' duration. Physical examination revealed oral candidiasis and lymphadenopathy. Patient 2, a man aged 36 years, presented with fever, cachexia, and cough and shortness of breath of 1 weeks' duration. Physical examination revealed lymphadenopathy. Additional laboratory studies on this patient revealed hematocrit 16% (normal: 39%-49%) and positive cerebrospinal fluid cryptococcal antigen. Both patients were treated with broad-spectrum antimicrobials for possible underlying bacterial infection and were discharged from the hospital.

Both patients had positive serology for HIV-1 antibody. Neither was receiving antiretroviral or antimycobacterial therapy. Patient 3, a man aged 28 years, presented with chronic fever and cough of 7 months' duration. Physical examination revealed cachexia and skin lesions. No lymphadenopathy was noted. Patient 4, a man aged 36 years, presented with fever, chronic fever, and diarrhea of 5 months' duration. Physical examination revealed oral candidiasis. No lymphadenopathy was detected. Both patients were treated with penicillin and chloramphenicol for underlying bacterial infection and were discharged from the hospital.

All four isolates were available for susceptibility testing. These isolates were resistant to all first-line drugs (isoniazid, rifampin, streptomycin, ethambutol, and pyrazinamide) used for treating M. tuberculosis infection and to alternative drugs (e.g., kanamycin and ciprofloxacin) used for treating atypical mycobacteria and multidrug-resistant tuberculosis (MDR-TB).

LB Reller, Clinical Microbiology Laboratory, Duke Univ Medical Center, Durham, North Carolina. LK Archibald, MD, WR Jarvis, MD, Div of Healthcare Quality Promotion; Div of AIDS, STD, and TB Laboratory Research, National Center for Infectious Diseases; LA Grohskopf, MD, EIS Officer, CDC.

Advances in laboratory methodology have enabled more rapid and reliable differentiation of mycobacterial species commonly associated with clinical illness (e.g., M. tuberculosis and MAC), and the identification of new or emerging species (e.g., M. triplex). However, ambiguities in determining specific mycobacteria species might occur in regions of the world where diagnostic resources are limited or not available. In addition, no standard susceptibility testing panel has been established for these organisms. These limitations might lead to difficulties in the clinical management of patients with disseminated mycobacterial infection.

The clinical manifestations of disseminated mycobacterial infection are nonspecific and are not indicative of the infecting species. Therefore, as with other mycobacterial infections, diagnosis and specific therapy should be guided by laboratory testing, including species identification and susceptibility testing whenever possible, rather than clinical findings alone.

The findings in this report are subject to at least three limitations. First, neither CD4 lymphocyte nor HIV-1 viral load data were obtained. However, because each patient had a marker of symptomatic HIV-1 infection (oral candidiasis, Kaposi's sarcoma, or positive cerebrospinal fluid cryptococcal antigen), all probably had clinical evidence of advanced immune deficiency. Second, because these patients had multiple conditions that could have produced their nonspecific symptoms and physical findings, it is unclear whether SAV mycobacteria were the cause of their symptoms. Further study and characterization of the SAV group of mycobacteria and of the clinical illness with which they are associated are required to better ascertain the prevalence and clinical significance of these mycobacterial infections. Finally, no information was available on treatment or postdischarge outcome for these patients.

Awareness of M. simiae and other SAV mycobacteria as potential causes of disseminated infection in patients with AIDS is important for several reasons. Because of the phenotypic similarity between SAV mycobacteria and other mycobacterial species, patients infected with SAV mycobacteria might go unrecognized and be presumed to be infected with other Mycobacterium species (e.g., M. tuberculosis), particularly in resource-poor settings without access to adequate laboratory testing. This might lead to ineffective treatment, because not all species are susceptible to all agents. Also, if these isolates were assumed to be M. tuberculosis, they could be misclassified as MDR-TB.

Because of the lack of data and of clinical experience with M. simiae and other SAV group mycobacteria, the best treatment is unknown. Infections with other mycobacteria, particularly M. tuberculosis, require treatment for prolonged periods with multiple agents to which the organisms are susceptible; not adhering to these principles promotes the development of drug-resistant organisms. Additional investigation is needed to determine whether similar hazards exist when SAV mycobacteria are treated with ineffective agents or otherwise suboptimal therapy.

This report is based on data contributed by S Tansuphasawadikul, B Eampokalap, A Chaovavanich, Bamrasnaradura Hospital, Nonthaburi; S Rheanpumikankit, Field Epidemiology Training Program, Ministry of Health, Thailand. P Kazembe, O Nwanyanwu, H Dobbie, Lilongwe Central Hospital, Lilongwe; Ministry of Health, Malawi. LF Turner, North Carolina Dept of Health and Human Svcs, State Laboratory of Public Health, Raleigh, North Carolina.

MMWR. 2002;51:497-501

2 figures omitted

In 2001, West Nile virus (WNV) activity was reported from 359 counties in 27 states and the District of Columbia (DC) to ArboNET, a web-based, surveillance data network maintained by 54 state and local public health agencies and CDC. This activity represented a marked increase from 2000, when WNV activity was reported from 138 counties in 12 states and DC.¹ This report summarizes surveillance data for 2001, which indicate that 66 human illnesses were reported from 10 states and that widespread WNV activity in birds, horses, and mosquitoes extended into the midwestern United States and several southern states unaffected previously. The findings in this report underscore the need for public education, increased WNV surveillance aimed at early viral detection, and sustained, integrated mosquito-control activities.

In 2001, CDC conducted WNV surveillance with 54 ArboNET surveillance coordinators from health departments in the contiguous 48 states and six jurisdictions (Chicago, DC, Houston, Los Angeles, New York City, and Philadelphia). Local WNV surveillance networks collected and tested for WNV or antibodies specimens from human and veterinary patients, dead birds, captive sentinel animals (mostly chickens), wild-caught birds, and mosquitoes. Test results, including county and week of specimen collection or illness onset, were entered into local electronic databases, and standardized summaries were forwarded weekly to CDC's ArboNET database system. In addition, reports of human WNV cases and other reports of WNV activity were reported to CDC by telephone, facsimile, or e-mail.

In 2001, a total of 66 human cases of WNV disease (64 persons with central nervous system infections [WNV meningoencephalitis] and two persons with uncomplicated WNV fever) were reported from 39 counties in 10 states. New York (13 WNV meningoencephalitis cases; two WNV fever cases), New Jersey (12 WNV meningoencephalitis cases), and Florida (12 WNV meningoencephalitis cases) accounted for 39 (59%) reported cases. Among 64 persons with WNV meningoencephalitis, the median age was 68 years (range: 9-90 years). Nine (14%) cases were fatal; the median age of these persons was 70 years (range: 44-90 years). The dates of human illness onset ranged from July 13 to December 7. In 36 (92%) counties reporting human cases, the first case was preceded by at least one report of a WNV-infected bird, sentinel animal, horse, or mosquito pool; 320 counties detected enzootic WNV activity but no human infections.

Of the 359 counties reporting WNV activity, 328 (91%) counties in 27 states and DC reported 7,333 dead WNV-infected birds (5,154 crows from two Corvus species, 966 blue jays, and 1,213 birds from 71 other avian species). In 238 (66%) counties, dead crows were the first indicators of WNV activity. Of 9,679 crows tested for WNV, 5,154 (53%) were positive for WNV infection compared with 2,179 (9%) of 24,898 birds from other species. Dead infected birds were collected during April 4–December 26. A total of 55 seropositive wild-caught birds were reported from DC and five counties in three states (Florida, New York, and Ohio) and represented the first detection of WNV activity in two of these counties. A total of 218 seroconverting captive sentinel animals were reported from 26 counties in five states (Florida, New Jersey, New York, North Carolina, and Virginia). In four Florida counties, seroconverting sentinel chickens were the first sign of WNV activity.

Horses were the only WNV-infected nonhuman mammals reported in 2001. A total of 733 equine cases were reported from 127 counties in 19 states (Alabama, Connecticut, Delaware, Florida, Georgia, Illinois, Indiana, Kentucky, Louisiana, Maryland, Massachusetts, Mississippi, New Hampshire, New Jersey, New York, North Carolina, Pennsylvania, Tennessee, and Virginia); this represented a 12-fold increase compared with 2000.¹ Florida reported 483 equine cases (66% of all reports) from 40 counties. The first equine illness preceded the first human illness; equine illness onset dates ranged from June 27 to December 18.

In 2001, a total of 564 counties conducted WNV testing on approximately 1.4 million mosquitoes from 91 species. WNV was detected in 919 mosquito pools (27 species) reported from 71 counties in 16 states (Connecticut, Delaware, Florida, Georgia, Illinois, Kentucky, Maryland, Massachusetts, Michigan, New Hampshire, New Jersey, New York, Ohio, Pennsylvania, Rhode Island, and Virginia) and DC. As in 2000, two enzootic vector species, Culex pipiens and Cx. restuans, collectively accounted for the majority (59%) of WNV-positive pools. WNV also was found for the first time in several additional species of potential public health importance, including Anopheles quadrimaculatus, Coquilletidia perturbans, Cx. nigripalpus, Cx. quinquefasciatus, Ochlerotatus sollicitans, Oc. taeniorhynchus, and Psorophora columbiae.

DR O'Leary, DVM, RS Nasci, PhD, GL Campbell, MD, AA Marfin, MD, Div of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, CDC.

The 2001 surveillance data indicate that the geographic area with WNV activity has increased and that dead WNV-infected birds were reported in western Arkansas, southern Maine, and southern Florida. Seven states reported human cases for the first time. Canadian health authorities also detected WNV activity in dead birds from southwestern Ontario, a region of lower latitude than the northern limits of WNV detection in the United States.² In 2001, a case of WNV encephalitis was diagnosed serologically in a resident of the Cayman Islands who had no recent travel history (CDC, unpublished data, 2001), suggesting that WNV has entered the Caribbean region. Human illness onsets on July 13 and December 7 in persons in Florida and Georgia, respectively, mark the earliest and latest reported human cases since the introduction of WNV to the United States. Extended seasonal activity in 2001 occurred in the northeast; two of five persons with illness onset on October 15 or later were from Massachusetts. The widespread occurrence of human cases and the occurrence of human cases outside of WNV's usual season (summer and early fall) suggest that (1) state and local health departments in the contiguous 48 states should, at a minimum, establish enhanced passive hospital-based surveillance for human cases of encephalitis of unknown etiology and (2) this surveillance should extend beyond mid-October.³

In 2001, infected birds, mosquitoes, or horses were detected in 16 states in which no previous WNV activity in animals had been reported. These findings demonstrate the dramatic spread of WNV westward and southward since 1999, when the virus was first recognized in North America. Although virus activity was detected for the first time in many southern states in 2001, the simultaneous appearance of two epizootic foci very early in 2001—one in the mid-Atlantic region and one in the southeast along the common borders of Florida, Georgia, and Alabama—suggests that WNV was introduced into the southern states by migrating birds in late 2000, but circulated at levels below the detection threshold of surveillance. Because many migratory bird species use well-established north-south flyways along the eastern seaboard, this movement of WNV from the mid-Atlantic region to the south Atlantic region and the Gulf states was expected; however, the reasons for WNV's rapid expansion into multiple foci in the central United States are less obvious. Possible mechanisms include carriage of the virus by the return of infected birds from wintering sites in southern states or by their incremental east-to-west local movements.⁴

Surveillance of dead birds is essential in monitoring WNV activity. Infection in species within the family Corvidae (e.g., crows and jays) is a particularly important indicator of WNV activity. In 2001, the proportion of tested birds that were infected continued to be disproportionately higher in crows than all other birds (53% versus 9%). Although 83% of infected birds reported were either crows or blue jays, this might be attributed to greater emphasis placed by states on monitoring these species. State and local health department surveillance programs should continue to emphasize the collection and testing of dead corvids. However, because noncorvid birds were first indicators of WNV activity in 57 (16%) of 359 counties where the virus was detected, surveillance programs should include these other species wherever possible. In 2001, serosurveillance of sentinel chickens and wild-caught birds contributed additional information on WNV transmission and provided collectively the initial signal of WNV activity in six counties. The limitations of these supplemental systems are documented^3,5 and their overall utility continues to be evaluated.

The 2001 equine WNV epizootic was unprecedented given its geographic span and the number of horses affected. In addition to a substantial epizootic in the northeast, an intense equine epizootic in Georgia and Florida accounted for 75% (551) of all reported equine cases. Scattered cases also were detected as far west as Louisiana, in the Ohio valley, and in northern Illinois. In August 2001, the U.S. Department of Agriculture granted conditional licensure of a commercial equine WNV vaccine because of the detrimental effect of these events on equine health and industry. Because WNV-infected horses are unlikely to develop viremias sufficient to infect feeding mosquitoes, they are unlikely to pose a risk to humans.⁶ However, equine epizootics reflect intense enzootic WNV activity in mosquitoes, which might place humans at increased risk.

In 2001, Culex mosquitoes (Cx. pipiens, Cx. restuans, and Cx. salinarius) were the most commonly identified mosquito vectors of WNV in the United States, and since 1999 these species have been found in close spatial and temporal proximity to the majority of human cases of WNV meningoencephalitis.^1,7- 8 Detection of WNV in several common human-feeding mosquito species (e.g., Cx. nigripalpus, Oc. sollicitans, Oc. taeniorhynchus, and Cq. perturbans) and recent studies demonstrating their ability to transmit this virus under laboratory conditions^9- 10 raise concerns about increased human risk in areas where these species are common.

The data available to the ArboNET system likely underestimate actual geographic distribution and intensity of WNV virus transmission in the United States. Data provided by the 54 ArboNET coordinators are derived largely from local health unit surveillance efforts, which vary according to capacity and ability. The 28 jurisdictions reporting activity probably support additional, undetected WNV transmission within their borders, and undetected foci of transmission probably exist in counties and states that have not reported transmission activity. In addition, some detected WNV infections might not have been entered into the ArboNET system.

In Florida, epizootic WNV activity has been reported since January 2002, indicating that year-round transmission is occurring in that state. In northern states, WNV activity has been reported since April. The extended seasonal activity, the broad vertebrate host and vector-mosquito range, and the establishment of multiple epizootic foci throughout the eastern United States demonstrate that WNV has established itself permanently in temperate North America and strongly suggest that it will spread further westward. This underscores the need for increased surveillance geared toward early viral detection and mosquito-control activities that weaken or break amplification cycles and decrease the risk for human and domestic animal infection with WNV. Prevention activities should continue to include (1) public education programs urging residential source reduction and personal protective measures to reduce mosquito exposure; (2) development of sustained, community-level integrated mosquito-surveillance and management programs³; and (3) high-priority emphasis on the control of urban Culex mosquitoes.

This report is based on data prepared by ArboNET surveillance coordinators in local and state health departments and ArboNET technical staff, Div of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, CDC.