contaminant of clinical specimens, the organism is generally viewed as a nonpathogenic. 2, 5–7 Yet, in recent years the frequency of true disease caused by nontuberculous mycobac- teria has increased, especially in immunocompromised hosts, 8 with reports of M. gordonae causing severe infections (e.g. disseminated disease 1, 9 –11 ) or pulmonary infection. 1, 12 These observations suggest that the clinical significance of M. gordonae isolation should not be underestimated in immuno- compromised hosts, especially in patients with AIDS. All case reports of M. gordonae bacteremia in AIDS patients were associated with disseminated disease accompanied by multi- organ system involvement and symptomatic disease. 9 –11 To the best of our knowledge this case represents the first report of sustained bacteremia without evidence of symptomatic disease. Although our patient was asymptomatic, we were com- pelled to treat his bacteremia because of his severe immuno- compromised state (CD4 + T cell count, 50/mm 3 ). The pa- tient was treated with three antimycobacterial drugs as recommended for the management of nontuberculous myco- bacterial bacteremia in severely immunocompromised HIV- infected patients. 13 However, this strategy reflects experience gathered from treatment of Mycobacterium avium complex infections and may not be optimal when treating bacteremia resulting from less pathogenic nontuberculous mycobacteria such as M. gordonae. We decided to initiate triple therapy to prevent symptomatic illness that typically accompanies dis- seminated M. gordonae infection. However, our treatment strategy may have some negative implications for the quality of life and medication adherence for this adolescent patient struggling with chronic illness. A less conservative approach such as initiating treatment and/or switching therapy to a single agent (i.e. clarithromycin or azithromycin), potentially risks the development of antimycobacterial drug resistance and increases the risk of dissemination to other organs. Robert Garofalo, M.D., M.P.H. Ellen G. Chadwick, M.D. Ram Yogev, M.D. Division of Infectious Diseases (RG, EGC, RY) and General Academic Pediatrics (RG) Children’s Memorial Hospital Chicago, IL Accepted for publication Feb. 20, 2003. Key words: Mycobacteria, bacteremia, adolescent, human im- munodeficiency virus/acquired immunodeficiency syndrome. Reprints not available. 1. Weinberger M, Berg SL, Feurstein IM, et al. Disseminated infection with Mycobacterium gordonae: report of a case and critical review of the literature. Clin Infect Dis 1992;14:1229 – 39. 2. Arnow PM, Bakir M, Thompson K, Bova J. Endemic contam- ination of clinical specimens by Mycobacterium gordonae. Clin Infect Dis 2000;31:472– 6. 3. Eckburg PB, Buadu EO, Stark P, et al. Clinical and chest radiographic findings among persons with sputum culture positive for Mycobacterium gordonae. Chest 2000;117:96 – 102. 4. Steere AC, Corrales J, von Graevenitz A. A cluster of Myco- bacterium gordonae isolates from bronchoscopy specimens. Am Rev Respir Dis 1979;120:214 – 6. 5. Peters M, Muller C, Rh sch-Gerdes S, et al. Isolation of atypical mycobacteria from tap water in hospitals and homes: is this a possible source of disseminated MAC infection in AIDS patients? J Infect 1995;31:39 – 44. 6. Good RC, Snider DE Jr. Isolation of nontuberculosis myco- bacteria in the United States, 1980. J Infect Dis 1982;146: 829 –33. 7. Tokars JI, McNeil M, Tablan O, et al. Mycobacterium gordo- nae pseudoinfection associated with a contaminated antimi- crobial solution. J Clin Microbiol 1990;28:2765–9. 8. Wolinski E. Mycobacterial diseases other than tuberculosis. Clin Infect Dis 1992;15:1–12. 9. Bernard E, Michiels JF, Pinier Y, et al. Disseminated infec- tion as a result of Mycobacterium gordonae in an AIDS patient. AIDS 1992;6:1217– 8. 10. Bonnet E, Massip P, Bauriaud R, et al. Disseminated Myco- bacterium gordonae infection in a patient with human immu- nodeficiency virus. Clin Infect Dis 1996;23:644 –5. 11. Chan J, McKitrick JC, Klein RS. Mycobacterium gordonae in the acquired immunodeficiency syndrome. Ann Intern Med 1984;101:400. 12. American Thoracic Society. Diagnosis and treatment of dis- ease caused by nontuberculous mycobacteria. Am J Respir Crit Care Med 1997;156(Suppl):S1–25. 13. Bartlett JG. The 2002 abbreviated guide to medical manage- ment of HIV infection. Baltimore: John Hopkins University School of Medicine, 2002:76. PARALYTIC POLIOMYELITIS CAUSED BY A VACCINE-DERIVED POLIO VIRUS IN AN ANTIBODY-DEFICIENT ARGENTINEAN CHILD We describe a case of poliomyelitis in a 3-year-old Argentinean boy with X-linked hypogammaglobu- linemia. The child had no history of polio vaccina- tion, but a poliovirus isolated from a stool sample had 97.2% genetic similarity to the Sabin 1 vaccine strain. According to the WHO definition, this is the first case reported of a vaccine-derived poliovirus infection (iVDPV) recorded in continental Latin America. The use of live oral polio vaccine (OPV) in the polio eradication initiative has led to elimination of indigenous wild polio and certification of the Americas as polio-free in September 1994. 1 An important feature of polioviruses in- cluding the Sabin strains in OPV is their continued genetic variation through mutation and recombination. Selective pressures in the human intestine can cause partial reversion of the attenuated OPV strains to neurovirulence, resulting in vaccine-associated paralytic poliomyelitis (VAPP), 2 which is clinically identical with the paralysis caused by wild virus, affecting both vaccine recipients and their contacts. Risk of VAPP is highest after the first dose of OPV, estimated as 1 case per 750 000 children vaccinated, and 3000 times higher in immunocompromised persons. Only OPV causes VAPP, the only indigenous form of polio in those countries that have eradicated wild poliovirus. 3, 4 A major challenge for the polio eradication “endgame” strategy is the recent characterization of vaccine-derived polio viruses (VDPV), defined by the WHO as a Sabin vaccine virus that shows 1% drift in the VP1 region of the genome. 5 Both types of VDPV, those circulating in the population (cVDPV) and those from immunodeficient long term excretors (iVDPV), may acquire both the neurovirulence and the trans- mission characteristics of wild-type strains, leading to polio outbreaks. In this communication we present one case of paralytic polio caused by an iVDPV in a primary antibody-deficient child and the implications for the polio eradication endgame strategy. Case report. The patient is a boy born in 1995 from consanguineous parents living in the rural area of the city of San Pedro, in the province of Buenos Aires, a community with 570 Vol. 22, No. 6, June 2003 THE PEDIATRIC INFECTIOUS DISEASE JOURNAL