Confirmed Mycoplasma pneumoniae Endocarditis

To the Editor: In Rosario, Argentina, during June 2005, a 15-year-old boy was hospitalized because of a 2-month history of fever. The patient had no history of cardiac disease or intravenous drug use. The results of the physical examination and the laboratory tests were within normal limits, except for an increased leukocyte count (14,000/µL) with 68% neutrophils. 
 
Transesophagic echocardiography showed mural vegetation on the right ventricle (30 mm × 20 mm) with no valve involvement. The patient was empirically treated with penicillin, gentamicin, and ceftriaxone. After treatment failed to produce a response, blood was submitted for culture for mycobacteria, brucellae, bartonellae, molds, and yeasts. BacT/ALERT bottles (bioMerieux, Durham, NC, USA), Hemoline performance biphasic medium (bioMerieux, Marcy L’Etoile, France), lysis centrifugation, and homemade culture broth were used. All culture results were negative. Results of PCR performed on serum for Actinobacillus actinomycetemcomitans were also negative. Because only the first samples were obtained before antimicrobial drug administration, a false-negative result was suspected. The patient underwent surgery for pulmonary microembolisms, and the vegetation was removed 4 weeks after drug treatment had started. The histologic appearance of the vegetation was consistent with infectious endocarditis, but the culture result was negative. 
 
After 6 weeks of treatment, the patient was discharged from the hospital; however, 10 days after discharge he again became febrile and was readmitted to the hospital. The vegetation was again found. On this second admission, all cultures were performed before administration of antimicrobial drugs, and several types of culture media were used. In the absence of any growth by day 6, the patient’s serum was screened for antibodies to Mycoplasma pneumoniae, Chlamydia pneumoniae, and Bartonella henselae. Serologic tests for immunoglobulin (Ig) G and IgM were conducted by indirect immunofluorescence assay (slides from Bion; Des Plaines, IL, USA) and fluorescein-labeled anti-human IgG and IgM (bioMerieux). For the IgM assay, the serum was pretreated with IgG/RF stripper (The Binding Site Ltd., Birmingham, UK). The titers for M. pneumoniae IgG and IgM antibodies were 2,048 and 160, respectively. Blood cultures were then subcultured in homemade Hayflick medium. These samples were incubated in 5% CO2 in a 37°C incubator and examined 2×/week for typical M. pneumoniae colonies. 
 
After 9 days of incubation, Hayflick agar plates inoculated with aliquots taken from homemade blood culture bottles (beef extract 5 g, yeast extract 5 g, peptone 10 g, glucose 2 g, NaCl 5 g, Na2HPO4 2.5 g, sodium heparin 10,000 U, distilled water to 1,000 mL, pH 7.6) showed colonies consistent with M. pneumoniae. No isolates were recovered from commercial blood culture bottles. 
 
Result of hemolysis test with sheep blood was positive. The isolate was definitively identified as M. pneumoniae after P1 cytadhesin gene amplification by nested PCR, with primers P1-40, P1-178, P1-285, and P1-331 (1). 
 
After mycoplasma were was isolated, intravenous clarithromycin was added to ceftriaxone; the ceftriaxone was discontinued 1 week later. The patient’s clinical condition improved, and he was discharged 3 weeks after bacteriologic diagnosis with a treatment regimen of oral levofloxacin. After 6 months of treatment, the vegetation was reduced with no evidence of calcification. 
 
Mycoplasma spp. have rarely been associated with endocarditis; until 2007, reports of only 8 cases had been published (2–8). The patient described herein had no underlying medical problems or immunodeficiency. Results of lymphocyte subsets, immunoglobulin titers, response to tetanus toxoid, and pneumococcal capsular polysaccharide were within reference ranges. 
 
Cases of culture-negative endocarditis are not routinely investigated for mycoplasmas; however, the role of these microorganisms as a cause of endocarditis might be underestimated. Mycoplasma spp. cannot be detected by Gram stain and are difficult to isolate in bacteriologic culture media. Commercial blood culture broths that use sodium polyanetholsulfonate as an anticlotting agent are not suitable for growing these microorganisms (9). Other diagnostic approaches include the detection of specific DNA sequences or the use of broad-range eubacterial primers in cardiac tissue (6). In the patient reported here, the clinical sample (vegetation) was not available for diagnostic M. pneumoniae gene amplification. We failed to detect M. pneumoniae by PCR-mediated gene amplification directly from whole blood and plasma. Theoretically, specific PCR should be more sensitive than culture, as shown in respiratory specimens, but to date attempts to detect M. pneumoniae in blood by PCR have not been successful. The bacterial load in blood may have been too low to detect the amplified product by ethidium bromide–stained gel electrophoresis. The larger volume of blood used and the preincubation in broth with yeast extract for 7 days could have improved the recovery by culture. Another cause of reduced PCR sensitivity may have been the use of frozen samples. 
 
This case of endocarditis caused by M. pneumoniae was confirmed by culture and occurred in a patient with no previous heart disease. Further studies are needed to evaluate the real incidence of M. pneumoniae as a cause of endocarditis as well as the occurrence of mycoplasma bacteremia in the absence of underlying infection of the endocardium.

drugs, and several types of culture media were used. In the absence of any growth by day 6, the patient's serum was screened for antibodies to Mycoplasma pneumoniae, Chlamydia pneumoniae, and Bartonella henselae. Serologic tests for immunoglobulin (Ig) G and IgM were conducted by indirect immunofl uorescence assay (slides from Bion; Des Plaines, IL, USA) and fl uorescein-labeled anti-human IgG and IgM (bioMérieux). For the IgM assay, the serum was pretreated with IgG/RF stripper (The Binding Site Ltd., Birmingham, UK). The titers for M. pneumoniae IgG and IgM antibodies were 2,048 and 160, respectively. Blood cultures were then subcultured in homemade Hayfl ick medium. These samples were incubated in 5% CO 2 in a 37°C incubator and examined 2×/week for typical M. pneumoniae colonies.
After 9 days of incubation, Hayfl ick agar plates inoculated with aliquots taken from homemade blood culture bottles (beef extract 5 g, yeast extract 5 g, peptone 10 g, glucose 2 g, NaCl 5 g, Na 2 HPO 4 2.5 g, sodium heparin 10,000 U, distilled water to 1,000 mL, pH 7.6) showed colonies consistent with M. pneumoniae. No isolates were recovered from commercial blood culture bottles.
After mycoplasma were was isolated, intravenous clarithromycin was added to ceftriaxone; the ceftriaxone was discontinued 1 week later. The patient's clinical condition improved, and he was discharged 3 weeks after bacteriologic diagnosis with a treatment regimen of oral levofl oxacin. After 6 months of treatment, the vegetation was reduced with no evidence of calcifi cation.
Cases of culture-negative endocarditis are not routinely investigated for mycoplasmas; however, the role of these microorganisms as a cause of endocarditis might be underestimated. Mycoplasma spp. cannot be detected by Gram stain and are diffi cult to isolate in bacteriologic culture media. Commercial blood culture broths that use sodium polyanetholsulfonate as an anticlotting agent are not suitable for growing these microorganisms (9). Other diagnostic approaches include the detection of specifi c DNA sequences or the use of broad-range eubacterial primers in cardiac tissue (6). In the patient reported here, the clinical sample (vegetation) was not available for diagnostic M. pneumoniae gene amplifi cation. We failed to detect M. pneumoniae by PCR-mediated gene amplifi cation directly from whole blood and plasma. Theoretically, specifi c PCR should be more sensitive than culture, as shown in respiratory specimens, but to date attempts to detect M. pneumoniae in blood by PCR have not been successful. The bacterial load in blood may have been too low to detect the amplifi ed product by ethidium bromide-stained gel electrophoresis. The larger volume of blood used and the preincubation in broth with yeast extract for 7 days could have improved the recovery by culture. Another cause of reduced PCR sensitivity may have been the use of frozen samples.
This case of endocarditis caused by M. pneumoniae was confi rmed by culture and occurred in a patient with no previous heart disease. Further studies are needed to evaluate the real incidence of M. pneumoniae as a cause of endocarditis as well as the occurrence of mycoplasma bacteremia in the absence of underlying infection of the endocardium.

Human Rhinovirus Group C Infection in Children with Lower Respiratory Tract Infection
To the Editor: Human rhinoviruses (HRVs), members of the family Picornaviridae, were fi rst isolated in 1956 (1); to date, >100 serotypes have been identifi ed on the basis of nucleotide sequence homologies. HRVs were previously divided into 2 genetic groups, HRV-A (n = 75) and HRV-B (n = 25). Recently, a putative new and distinct rhinovirus group, HRV-C, has been reportedly found in some patients with respiratory tract infections (RTIs) (2)(3)(4)(5)(6)(7)(8). To extend these initial fi ndings and assess the pathogenicity of HRV-C, we investigated its prevalence as well as its clinical and molecular features in children with lower acute RTIs in Beijing, People's Republic of China.
From July through December 2007, nasopharyngeal aspirates were collected from 258 children (167 boys and 91 girls) who had lower acute RTIs at the time of their admission to Beijing Children's Hospital. The children were 1 month to 15 years of age (mean age 37 months, median age 10 months). Nucleic acids were extracted from clinical samples by using the NucliSens easyMAG platform (bioMérieux SA, Marcy L'Etoile, France). Each specimen was tested for the presence of common respiratory viruses: human parainfl uenza viruses 1-4, infl uenza viruses, respiratory syncytial virus, enteroviruses, human coronaviruses (229E, NL63, HKU1, and OC43), metapneumovirus, adenoviruses, and bocaviruses. To study the prevalence of HRV-C, we designed a specifi c reverse transcription-PCR (RT-PCR) that generated a 330-bp PCR product encompassing a portion of the 5′-untranslated region, the full virus capsid protein (VP) 4 gene, and a portion of the VP2 gene of the HRV-C genome. (All primer sequences and protocols of these assays are available from J.W. upon request.) This RT-PCR detected HRV-C in 14 patients (12 boys and 2 girls, 1 month to 13 years of age [mean age 19 months, median age 6 months]). In 6 of the 14 patients, HRV-C was the only virus detected, which suggests a direct correlation between HRV-C infection and lower acute RTIs. In the remaining 8 patients, other respiratory viruses were also detected. Respiratory syncytial virus, the most important cause of lower acute RTIs in children, was codetected in 7 of the HRV-Cpositive patients, and human parainfl uenza virus 3 was codetected in the other patient. Human coronavirus NL63 was codetected with respiratory syncytial virus in 1 HRV-C-positive patient.
HRV-C infection may be seasonal. This virus was detected during only 3 of the 6 months in which specimens were collected. Specifi cally, HRV-C was detected in samples collected in October (7/50), November (5/96), and December (2/8) but not in those collected in July (0/37), August (0/42), or September (0/25). In contrast, HRV-A and HRV-B were detected in