Jpn. J. Infect. Dis., 66, 475-479, 2013

Original Article simiae: a Possible Emerging Pathogen in Iran

Abdolrazagh Hashemi-Shahraki1, Davood Darban-Sarokhalil2*, Parvin Heidarieh2, Mohammad Mehdi Feizabadi3, Shima Deshmir-Salameh1, Sedigheh Khazaee4, and Seyed Mohammad Alavi1 1Infectious and Tropical Diseases Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz; 2Department of Bacteriology & Virology, School of Medicine, Alborz University of Medical Sciences, Karaj; 3Department of Microbiology, School of Medicine, Tehran University of Medical Sciences, Tehran; and 4Molecular Pathology Research Center, School of Medicine, Kermanshah University of Medical Sciences, Emam Reza Hospital, Kermanshah, Iran (Received February 25, 2013. Accepted June 25, 2013)

SUMMARY: Mycobacterium simiae has been reported worldwide, particularly from the Middle East. This organism has been recognized as a causative agent of pulmonary and disseminated infections. In this study, we used molecular methods to detect this organism from patients who were suspected of having tuberculosis. A total of 117 isolates of mycobacteria were evaluated from different regions of Iran. Isolates were identified using phenotypic methods and gene sequencing of 16S rRNA, rpoB, hsp65, and ITS. Of the 117 isolates, 12 M. simiae isolates (10.2z) were identified from different clinical samples, including bronchoalveolar lavage and sputum (n ＝ 8), blood (n ＝ 3), and lymph node biopsy (n ＝ 1). Three isolates (3/12, 25z) were recovered from blood samples of HIV cases when the CD4＋ count was less than 50/ml. There was no significant relationship between infection and age or gender. Infection with nontuberculous mycobacteria (NTM), including M. simiae,isthemajorproblemamong immunocompromised patients. The results of this study illustrated the importance of molecular methods for accurate and rapid detection of NTM infections in the treatment of nonresponding patients with suspected tuberculosis.

newer drugs such as 8-methoxy fluoroquinolone (mox- INTRODUCTION ifloxacin), clarithromycin, and trimethoprim/sulfame- Mycobacterium simiae was reportedly restricted to thoxazole are effective on this mycobacterium. There- several regions of the world such as Cuba, Gaza, and fore, rapid detection of M. simiae, particularly among southern United States (1–5). Recently, this organism immunocompromised patients, is necessary to prescribe has been widely reported from other areas, including the an appropriate therapy regimen. Middle East (6–9). Transmission of M. simiae to hu- Using conventional bacteriology, it is difficult to de- mans and possible environmental sources of this organ- tect and identify M. simiae, particularly when the result ism remain unclear; however, the routes of entry are of the niacin test is positive (1). Due to the progress in probably respiratory and digestive tracts (10). molecular genetic tools, it has now become possible to In the M. simiae complex, M. simiae is clinically most detect NTM species (11). relevant. M. simiae is one of the most common non- M. simiae has been previously reported from Iran tuberculous mycobacteria (NTM) causing lung disease. (8,9). In the present study, we attempted to detect this It causes pulmonary disease in the patients with under- organism from patients diagnosed with TB using lying disease and disseminated infection in both im- molecular methods. These patients were empirically munocompromised and immunocompetent patients treated with anti-TB drugs routinely prescribed to acid- (1,4). fast bacillus (AFB) sputum smear-positive pulmonary There is limited data on the association between the in patients. To the best of our knowledge, this is the first vitro susceptibility of the organism and the in vivo study regarding M. simiae in Ahvaz, Kermanshah, and response to most drugs (1). M. simiae usually shows Tehran. poor in vivo response to therapy, and most isolates are resistant to first line anti-tuberculosis (TB) drugs such as MATERIALS AND METHODS isoniazid and rifampicin (1,7). The therapy regimen for this pathogen completely differs from that for TB, and A total of 117 clinical strains of mycobacteria were isolated from 190 cases with suspected Mycobacterium- *Corresponding author: Mailing address: Department of related complications at the TB Reference Centre of Bacteriology & Virology, School of Medicine, Alborz Ahvaz (Khuzestan, Iran), Masoud Laboratory (Tehran, University of Medical Sciences, Karaj, Iran. Tel: ＋98- Iran), and Kermanshah (Kermanshah, Iran) between 26-34336007, Fax: ＋98-26-34319188, E-mail: davood_ 2009 and 2012 using standard decontamination proce- darban＠yahoo.com dures, acid-fast staining, and culture on L äowenstein–

475 Jensen (LJ) medium (12). All strains were stored at the sequences of mycobacteria were separately retrieved TB Reference Centre of Ahvaz at －709C in appropriate from the GenBank database (www.ncbi.nlm.nih.gov/ medium consisting of tryptic soy broth supplemented genbank) and deposited in the jPhydit program as an in- with 20z (v/v) glycerol. Based on the clinical findings house-prepared database of 16S rRNA, rpoB, hsp65, and AFB staining results, all the patients were subjected and ITS sequences (17). The obtained sequences of to the routine anti-TB treatment protocol. different genes for each clinical isolate were aligned and Of the 190 cases, 57, 98, and 35 cases were registered compared to determine the percentage of similarity us- as TB cases from Kermanshah, Khuzestan, and Tehran, ing the jPhydit program (17). Phylogenetic trees were respectively. The strains were isolated from various clin- constructed from DNA sequences using the neighbor- ical samples, including bronchoalveolar lavage (BAL), joining (NJ) method and Kimura's 2-parameter (K2P) sputum, abscess, blood, gastric washing, urine, soft tis- distance correction model with 1000 boot-strap replica- sue infection, and lymph node biopsy. From each case tions supported by the MEGA 4.1 software (18). with pulmonary disease, 3 independent sputum, gastric The GenBank accession numbers of the clinical iso- washing, or BAL samples were examined. Three soft tis- late AFP-000175 as a representative strain of M. simiae sue infection and 2 biopsy samples from abscess and strain determined in this study are as follows: JX266703 lymph node, respectively, were examined for the isola- for 16S rRNA, JX294386 for hsp65, JX294404 for tion and identification of mycobacteria. From cases rpoB, and JX270663 for ITS. with disseminated disease, 3 blood samples were cul- tured. All the isolated mycobacterial strains were identi- RESULTS fied to the species level using conventional tests (12). Of the 117 strains of mycobacteria, 12 were positive for Of the 190 TB cases, a total of 117 strains of niacin production, which is the most well-known pheno- mycobacteria were recovered from various clinical sam- typic characteristic of Mycobacterium tuberculosis; ples (Fig. 1). Based on phenotypic features, the species however, all the other strains produced yellow-pigment- spectrum consisted of 94 (80.3z) M. tuberculosis,12 ed colonies as well as showed poor (10.2z) M. simiae-like, 5 (4.3z) M. kansasii,and6 response to TB treatment. For reliable identification, 12 (5.2z) M. fortuitum. The isolates belonging to M. slow-growing mycobacteria were subjected to further tuberculosis, M. fortuitum,andM. kansasii were clear- identification using a battery of molecular tests. ly identified at the species level by phenotypic analysis. Amplification and sequencing of 16S rRNA, rpoB, Of the 12 M. simiae-like strains, optimum growth was hsp65, and ITS: (i) 16S rRNA: The 16S rRNA genes observed at 379C, with colonies appearing after 4 from isolates were amplified using primers, pA (5?- weeks. No growth was observed at 429C. On Mid- AGAGTTTGATCCTGGCTCAG-3?) and pH (5?- dlebrook 7H10 agar medium, 12 strains were AAGGAGGTGATCCAGCCGCA-3?), as described scotochromogenic with yellow pigmentation. Twelve, previously (13). Genomic DNA of M. tuberculosis 10, and 8 isolates were positive for niacin, tellurite H37RvT and double-distilled water were used as positive reduction, and urease activity, respectively, and all were and negative control, respectively, in all PCR analyses. positive for semi-quantitative catalase À45 mm and The PCR-amplified 16S rRNA gene products for each isolate were purified using the GeneJETTM Gel Extrac- tion Kit (Fermentas, Ukraine), according to the manufacturer's instructions. The sequences of the products were determined using the ABI PRISM} 7700 Sequence Detection System (Applied Biosystems, Foster City, Calif., USA) following the standard protocol of the supplier. (ii) hsp65: A 439-bp amplified PCR product of the hsp65 gene for each isolate was purified and sequenced, as described above using 2 specific primers, Tb11 (5?- ACCAACGATGGTGTGTCCAT-3?)andTb12(5?- CTTGTCGAACCGCATACCCT-3?)(14). (iii) rpoB: A 750-bp fragment of the rpoB gene was amplified and sequenced using 2 specific primers, MycoF (5?-GGCAAGGTCACCCCGAAGGG-3?)and MycoR (5?-AGCGGCTGCTGGGTGATCATC-3?), as described previously (15). (iv) ITS: The universal primers, L (5?-GCTGGAT CACCTCCTTCT-3?) from the conserved sequence at the 3? endofthe16SrRNA(fromposition1525toposi- tion 1543 on the Escherichia coli 16S rRNA) and R (5?- Fig. 1. The distribution of M. simiae isolates in three provinces CTGGTGCCAAGGCATCCA-3?), which was deduced of Iran. Khuzestan: 98 tuberculosis (TB) cases (63 cultures from the conserved 5? sequence of 23S rRNA (position positive); M. fortuitum (n ＝ 4), M. kansasii (n ＝ 2), M. 23 to position 40 on the Escherichia coli 23S rRNA), simiae (n ＝ 4), M. tuberculosis (n ＝ 53). Tehran: 35 TB cases (23 cultures positive); M. fortuitum (n ＝ 1), M. kansasii (n ＝ were used for amplification of the full length of the ITS 2), M. simiae (n ＝ 3), M. tuberculosis (n ＝ 17). Kermanshah: region (16). 57 TB cases (31 cultures positive); M. fortuitum (n ＝ 1), M. Analysis of sequence data: All existing relevant kansasii (n ＝ 1), M. simiae (n ＝ 5), M. tuberculosis (n ＝ 24).

476 negative for 14-day arylsulfatase. The results obtained Table 1 shows the baseline characteristics and symp- in this study were sufficient to identify the strains as M. toms of the cases. Three strains (3/12, 25z)wererecov- simiae-like strains, mainly because they produced niacin ered from blood samples of HIV cases when the CD4＋ and were scotochromogenic with yellow pigmentation. count was less than 50/ml.Foreachcaseofdisseminated Twelve M. simiae-like strains were isolated on pure M. simiae, the most common symptoms were fever, culture from BAL and sputum (n ＝ 8), blood (n ＝ 3), night sweats, anorexia, and weight loss. The mean age and lymph node biopsy (n ＝ 1). All cases with pulmo- of the 6 female and 6 male patients was 57.1 and 48.3 nary disease were symptomatic, and 3 independent BAL years, respectively. specimens from each case were AFB-positive and yield- ed a pure culture of the same strains of mycobacteria. DISCUSSION From cases with disseminated disease, 3 different blood samples were cultured, which leaded to the isolation of M. simiae is commonly isolated from respiratory few colonies of mycobacteria. AFB staining was nega- specimens (4). Previously, this organism was presumed tive for all isolates from all 3 cases with disseminated to cause disseminated infections only in immunocom- disease. Slow-growing scotochromogenic mycobacteria promised individuals only. However, recent reports from specimens that were obtained from 2 independent have confirmed that M. simiae can also disseminate in fine-needle lymph node biopsies of a patient suffering immunocompetent patients (10,19–21). The rate of from cervical lymphadenitis and grown on LJ medium NTM infections varies geographically worldwide. M. exhibited positive AFB staining. The etiological role of simiae has been reported from different parts of Asia, the 12 isolates may be inferred from the fact that AFB including Iran, Japan, and Turkey (8,9,22,23). This or- were microscopically observed in all BAL specimens ganism may have been an infectious agent in Iran. A and 1 biopsy sample, and the same strains of mycobac- poor laboratory diagnostic system is assumed to be the teria were recovered in pure culture from different clini- reason for failure in detecting this organism from cal samples of each patient according to the American patients infected with M. simiae. Thoracic Society (ATS) guidelines (1). The correlation between M. simiae infection and age The respective 16S rRNA (1426 bp), rpoB (661 bp), and gender has remained controversial, although in- hsp65 (413 bp), and ITS (214 bp) genes from each iso- fected patients tended to be older than those infected late were identical. Comparison of the 16S rRNA, rpoB, with TB (7). The rate of M. simiae infections among hsp65, and ITS sequences from each isolate using our females varies from low to high among studies (5,7). in-house database in the jPhydit program revealed that Themeanageofthepatientswas58.2± 16.9 years. M. simiae strain ATCC 25275T was the best match for Our findings regarding age are in agreement with each sequence. Phylogenetic trees of 16S rRNA (Fig. 2) other reports from Iran (8). and other gene (rpoB, hsp65, and ITS) sequences (data ThetrueinfectionrateforM. simiae is reportedly not shown) also confirmed the identification of the iso- 9–24z (7,18), which is higher than the rate determined lates as M. simiae. Based on these data, we concluded in the present study. A previous study reported that the that the true identity of the causative microorganism most frequently isolated NTM were M. simiae (7). NTM was M. simiae. isolated in the present study included 12 M. simiae,5M. Based on current data, the rates of M. simiae infec- kansasii,and6M. fortuitum isolates, consistent with tions in Kermanshah, Khuzestan, and Tehran were the ATS diagnostic criteria and were considered clinical- 16.1z (5/31 cultures positive), 6.3z (4/63 cultures ly relevant. However, due to the small number of cases, positive), and 13z (3/23 cultures positive), respectively. the true incidence of NTM infections is likely underesti-

Fig. 2. Phylogenetic tree of the genus Mycobacterium computed from the 16S rRNA gene sequences (1426 bp) of 12 Iranian clinical isolates of M. simiae by the NJ method and K2P distance correction model with 1000 bootstrap replications supported by the MEGA 4.1 software. Percentages at nodes represent bootstrap levels supported by 1000 resampled datasets. Bar 0.005 substitutions per nucleotide position.

477 Alive Dead Outcome 6) — identification Treatment after correct CLR, RIF (3 months) Alive CLR, EMB, CIP (5 months) Alive CLR, EMB, CIP (1 month) Dead 5) hritis. — — — Initial TB/RIF, INH CLR, RIF (3 months) Alive TB/RIF, INH BRFIHCREBCP3ek)Dead CLR,EMB,CIP(3weeks) TB/RIF,INH TB/RIF, INHTB/RIF, INH CLR, MOX (2 months) CLR, RIF (3 months) Alive Alive TB/RIF, INH No available BRFIHCREBCP2ots Alive CLR,EMB,CIP(2months) TB/RIF,INH M. simiae diagnosis/treatment — — — — Chest X-ray Cavity in left lower lob Cavity in right middle lobe Pleural effusion TB/RIF, INH CLR, RIF (3 months) Alive 4) — — Symptom CH,LFvDs Infiltrates PC, WL, Fev, DyspPC, WL, Fev, Dysp Infiltrates PC,Hp,WL,Fev,Dysp PC, Infiltrates WL, Fev, DyspPC,Hp,WL,Fev,Dysp Infiltrates Pleural effusionPC,Hp,WL,Fev,Dysp TB/RIF, INHPC,Hp,WL,Fev,MA RIF, INH,PC, CLR Hp, (6 WL months)Fev, WL Alive PC, Hp, WL ＋ ＋ ＋ ＋ ＋ － － ＋ ＋ － ＋ － Direct microscopy 3) Sample source Sputum Lymph node biopsy Blood Blood 2) Table 1. Baseline characteristics, symptoms, and outcome of the patients infected by — BAL history Post medical Previous TB, AIDS Sex (yr) Age 1) T 71 Female Bronchiectasis BAL K 60 Female — BAL K 15 Female — BAL K 62 Male Malignancy BAL K 60 Male Kh 89 Female RA Kh 31 Male AIDS Kh 45 Male AIDS, DM Blood Isolate Region : K, Kermanshah; T,: Tehran; DM, Kh, diabetes Khuzestan. mellitus: type BAL, II; bronchoalveolar ILD, lavage interstitial: lung fluid. PC, disease; productive COPD, cough; chronic: Hp, obstructive RIF, hemoptysis; pulmonary rifampicin; WL, disease; INH, weight: TB, isoniazid. loss; CLR, tuberculosis; MA, clarithromycin; RA, malaise/asthenia; EMB, rheumatoid Dysp, ethambutol; art dyspnea; CIP, Fev, ciprofloxacin; fever. MOX, moxifloxacin. 1) 2) 3) 4) 5) 6) K12 Kh4 Kh29 K7 K 80 Male — BAL Kh7 T11Kh19 TWK12 Kh 40 12 Female Male COPD BAL K18 K27 AFP-000175 T 68 Female ILD K13

478 mated. In future, a large-scale prospective cohort study 4. Balkis, M.M., Kattar, M.M., Araj, G.F., et al. (2009): Fatal dis- is required to explore the incidence rate of NTM seminated Mycobacterium simiae infection in a non-HIV patient. Int. J. Infect. Dis., 13, 286–287. infections. 5. van Ingen, J., Boeree, M.J., Dekhuijzen, P.N., et al. (2008): Similar to a previous study (20), productive cough, Clinical relevance of Mycobacterium simiae in pulmonary sam- dyspnea, malaise, fever, sweat, and body weight loss ples. Eur. Respir. J., 31, 106–109. were the most common signs and symptoms of M. 6. Shitrit, D., Peled, N., Bishara, J., et al. (2008): Clinical and simiae infection among the Iranian patients. However, radiological features of Mycobacterium kansasii infection and Mycobacterium simiae infection. Respir. Med., 102, 1598–1603. the most frequent imaging finding was nodular lesions 7. Maoz, C., Shitrit, D., Samra, Z., et al. (2008): Pulmonary (100z) in Iranian patients, which makes our study Mycobacterium simiae infection: comparison with pulmonary distinct from those in a previous study (8). The reason tuberculosis. Eur. J. Clin. Microbiol. Infect. Dis., 27, 945–950. for such difference remains unclear. We believe that 8. Baghaei, P., Tabarsi, P., Farnia, P., et al. (2012): Pulmonary dis- socioeconomic and geographic differences may be con- ease caused by Mycobacterium simiae in Iran's national referral center for tuberculosis. J. Infect. Dev. Ctries., 6, 23–28. sidered; however, to arrive at a definite conclusion, 9. Mirsaeidi, S.M., Tabarsi, P., Mardanloo, A., et al. (2006): Pul- future study is recommended. monary Mycobacterium simiae infection and HTLV1 infection: Three patients in our study were HIV- and M. simiae- an incidental co-infection or a predisposing factor? Monaldi positive; however, the rate of M. simiae infection, par- Arch. Chest Dis., 65, 106–109. 10. Narang, R., Narang, P., Jain, A.P., et al. (2010): Disseminated ticularly in non-HIV patients, remains vague (22). In an disease caused by Mycobacterium simiae in AIDS patients: a earlier study conducted in Iran, all patients, except 1, report of three cases. Clin. Microbiol. Infect., 16, 912–914. were HIV-negative (8). Thus, we think that further stu- 11. Pai, S., Esen, N., Pan, X., et al. (1997): Routine rapid Mycobac- dies on a larger scale are warranted to clarify this associ- terium species assignment based on species-specific allelic varia- tion in the 65-kilodalton heat shock protein gene (hsp65). Arch. ation. Pathol. Lab. Med., 121, 859–864. Unlike M. tuberculosis, M. simiae is not an obligate 12. Kent, P.T. and Kubica, G.P. (1985): Public health mycobacteriol- pathogen. Therefore, it normally does not cause disease ogy: a guide for the level III laboratory. Centers for Disease Con- in immunocompetent individuals (1). It mainly causes trol, U.S. Department of Health and Human Services, Atlanta, soft-tissue infections, lymphadenitis, and lung diseases Ga. 13. Edwards, U., Rogall, T., Bl äocker, H., et al. (1989): Isolation and and occasionally causes disseminated infection in direct complete nucleotide determination of entire genes. Charac- patients with impaired immune defenses, such as AIDS terization of a gene coding for 16S ribosomal RNA. Nucleic Acids (1). In the present study, one-fourth of M. simiae- Res., 17, 7843–7853. infected patients had disseminated disease; all of them 14. Telenti, A., Marchesi, F., Balz, M., et al. (1993): Rapid identifi- were HIV-positive and had low CD4＋ counts. This find- cation of mycobacteria to the species level by polymerase chain reaction and restriction enzyme analysis. J. Clin. Microbiol., 31, ing is in agreement with those of previous studies 175–178. (7,24,25). The clinical presentation, radiographic find- 15. Adekambi, T., Colson, P. and Drancourt, M. (2003): rpoB-based ings, and utility of sputum-based diagnostic modalities identification of nonpigmented and late-pigmenting rapidly grow- may differ between HIV-positive and -negative cases ing mycobacteria. J. Clin. Microbiol., 41, 5699–5708. 16. Leblond-Bourget, N., Philippe, H., Mangin, I., et al. (1996): 16S and may differ in HIV-infected individuals as immune rRNA and 16S to 23S internal transcribed spacer sequence ana- suppression progresses (1,26). These factors increase the lyses reveal inter- and intraspecific Bifidobacterium phylogeny. complexity of management of patients with HIV/M. Int. J. Syst. Bacteriol., 46, 102–111. simiae coinfections. 17. Jeon, Y.S., Chung, H., Park, S., et al. (2005): jPHYDIT: a All patients in the present study were diagnosed with JAVA-based integrated environment for molecular phylogeny of ribosomal RNA sequences. Bioinformatics, 21, 3171–3173. TB and received treatment for several years (repeated 18. Tamura, K., Dudley, J., Nei, M., et al. (2007): MEGA4: Molecu- courses of anti-TB drugs) without showing any response lar Evolutionary Genetics Analysis (MEGA) software version 4.0. to anti-TB drugs. These data highlighted the importance Mol. Biol. Evol., 24, 1596–1599. of molecular methods for the accurate and rapid detec- 19. Piquero, J., Casals, V.P., Higuera, E.L., et al. (2004): Iatrogenic Mycobacterium simiae skin infection in an immunocompetent tion of NTM infections, particularly those caused by M. patient. Emerg. Infect. Dis., 10, 969–970. simiae, and the importance of choosing an optimal 20. Chung, K.P., Tsai, T.H., Huang, Y.T., et al. (2009): Refractory treatment for nonresponding TB patients. bacteremia and osteomyelitis resulting in fatal bacteremic pneu- This study, as well as previously published reports, monia with multiorgan failure caused by Mycobacterium simiae in a non-human immunodeficiency virus-infected adult. J. Clin. emphasized the fact that NTM, such as M. simiae,are Microbiol., 47, 3777–3779. mainly problematic in an immunocompromised state 21. Amorim, A., Macedo, R., Lopes, A., et al. (2010): Non-tuber- for either HIV-positive or -negative patients. culous mycobacteria in HIV-negative patients with pulmonary disease in Lisbon, Portugal. Scand. J. Infect. Dis., 42, 626–628. 22. Yoshimura, K., Imao, M., Goto, H., et al. (2005): A case of pul- Conflict of interest None to declare. monary infection due to Mycobacterium simiae. Nihon Kokyuki Gakkai Zasshi, 43, 32–36 (in Japanese). 23. Onen, Z.P., Karahan, Z.C., AkkocaYƒldƒz, O., et al. (2010): REFERENCES Mycobacterium simiae infection in an immunocompetent patient, 1. Griffith, D.E., Aksamit, T., Brown-Elliot, B.A., et al. (2007): An with DNA analyses verification. Tuberk. Toraks., 58, 306–310. official ATS/IDSA statement: diagnosis, treatment, and preven- 24. Matveychuk, A., Fuks, L., Priess, R., et al. (2012): Clinical and tion of nontuberculous mycobacterial diseases. Am. J. Respir. radiological features of Mycobacterium kansasii and other NTM Crit. Care Med., 175, 367–416. infections. Respir. Med., 106, 1472–1477. 2. Huminer, D., Dux, S., Samra, Z., et al. (1993): Mycobacterium 25. El-Solh, A.A., Nopper, J., Abdul-Khoudoud, M.R., et al. simiae infection in Israeli patients with AIDS. Clin. Infect. Dis., (1998): Clinical and radiographic manifestations of uncommon 17, 508–509. pulmonary nontuberculous mycobacterial disease in AIDS 3. El Sahly, H.M., Septimus, E., Soini, H., et al. (2002): Mycobac- patients. Chest, 114, 138–145. terium simiae pseudo-outbreak resulting from a contaminated 26. Sampaio, J.L., Pereira, R.M., Souza, J.R., et al. (2001): hospital water supply in Houston, Texas. Clin. Infect. Dis., 35, Mycobacterium simiae infectioninapatientwithacquiredim- 802–807. munodeficiency syndrome. Braz. J. Infect. Dis., 5, 352–355.

479479