Mycoplasma hominis is a small prokaryotic organism belonging to the genus Mycoplasma. This organism lacks a cell wall, rendering it undetectable by Gram staining and making it resistant to beta-lactam antibiotics. Unlike other Mycoplasma species, M. hominis can grow on blood agar and chocolate agar plates; however, such growth is very slow and special techniques are required to identify this organism. Thus, it is very difficult to diagnose infections caused by this pathogen.⁴ M. hominis most commonly colonizes the urogenital tracts of sexually active adults and can be a causative agent of their urogenital tract infections. In pediatric populations, this organism is sometimes associated with invasive neonatal infections^5,8,20,22 due to the acquisition of the microbe from a colonized birth canal, although asymptomatic colonization has also been reported.^16–19 However, invasive infections by M. hominis after the neonatal period are very rare. In the past, this organism has been reported as a potential cause of some severe childhood infections following the neonatal period,^6,7,9,13 including a few invasive cases with central nervous system (CNS) involvement.^10,21 Although previous reports have suggested that immunosuppression may play a role in cases of extragenital infection caused by M. hominis,^11,21 there are no known risk factors for M. hominis infection and immunocompetent children can be infected by the organism. Because of the low number of reported cases, the standard antibiotic regimen for CNS infection caused by M. hominis is undetermined. In addition, there have been only a few reports of device-related infections associated with this organism and the optimal regimen for the management of such infections remains unknown.

Here, we present a successfully treated case of ventriculoperitoneal shunt (VPS) infection with CNS involvement caused by M. hominis.

Case Report

A 6-year-old girl with hydrocephalus resulting from aqueductal stenosis underwent VPS insertion. She had undergone endoscopic third ventriculostomy (ETV) 6 months previously; however, follow-up MRI revealed re-obstruction of her third ventricle floor. Because fenestration of her third ventricle floor on the first ETV had been difficult due to its anatomical limitations, repeat ETV was not considered. The patient also had a diagnosis of pan-hypopituitarism and was taking hydrocortisone at 0.25 mg/kg/day. Cefazolin was administered for prophylaxis of surgical site infection (SSI) until her 5th postoperative day. Although transient fever was observed for a few days during her hospitalization, no symptoms of the SSI were observed and the fever subsided without treatment. The patient appeared to have recovered completely and was subsequently discharged on her 8th postoperative day. However, 2 days after discharge, she returned with intermittent fever and was readmitted into our hospital.

Upon readmission, the patient had a fever of 38.4°C but was alert, with good tissue perfusion and without headache. The abdominal incision was erythematous (Fig. 1), and the skin overlying the VPS on the left side of the patient's head was slightly swollen. Her abdomen was soft and flat. A sepsis workup was performed, including blood and CSF cultures. Laboratory data upon her readmission were as follows: leukocytes, 16.5 × 10⁹/L; C-reactive protein, 48.7 mg/L; and blood glucose, 5.5 mmol/L. Other values remained normal. CSF analysis revealed 33.0 × 10⁶/L leukocytes with 31% polynuclear cells, a protein value of 90.0 mg/L, and a glucose level of 3.4 mmol/L. Although no microorganisms were detected by Gram staining of the CSF, treatment with intravenous vancomycin plus ceftazidime was empirically started for suspected meningitis and VPS infection. On the same day, the VPS was initially externalized in the operating theater. At that time, a small amount of pus was drained from the abdominal incision. Gram staining of the pus also failed to reveal any microorganisms.

Image showing the patient's abdomen with redness around the incision site. Figure is available in color online only.

The patient remained febrile for 3 days despite the administration of antibiotics and externalization of the VPS, but blood and CSF cultures performed upon her readmission appeared to be negative. On her 4th day in the hospital, however, the sample of pus obtained from the abdominal incision grew pinpoint colonies on both blood agar and chocolate agar plates (Fig. 2). Because Gram staining of the colonies failed to reveal microorganisms, atypical organisms were assumed to be causative pathogens. We thus used 16S ribosomal RNA sequencing to identify the isolate responsible, as previously described.¹² Eventually, the isolate was identified as M. hominis. Subsequently, the intravenous antibiotics were changed to ciprofloxacin (10 mg/kg every 12 hours) plus clindamycin (13 mg/kg every 8 hours) on the patient's 6th day in the hospital. Soon after this change of the antibiotic regimen, the fever subsided. On the patient's 11th day in the hospital, the CSF sample obtained upon her readmission also grew colonies of M. hominis, indicating that the organism had also caused meningitis. Antibiotic administration was continued and CSF and serum were monitored for drug concentrations. Cultures and polymerase chain reaction (PCR) assays, using specific primers for M. hominis as previously described,³ were repeated for CSF samples on the patient's 5th, 8th, 11th, and 15th days in the hospital. Since culture and PCR results continued to be negative following the change in antibiotic regimen, the externalized VPS was removed and a new temporary external ventricular drain (EVD) was inserted on her 18th day in the hospital. Antibiotics were continued after the operation. Cultures and PCRs on CSF samples repeated every few days (on the patient's 22nd, 25th, 29th, 33rd, and 37th days in the hospital) consistently showed negative results. On the patient's 38th day in the hospital, a new VPS was inserted and the temporary EVD was removed. Antibiotic treatment was continued for an additional week following insertion of the new VPS. Subsequently, the patient had an uncomplicated postoperative course. CSF cultures and PCR assays for M. hominis at the time of the final surgical intervention were both negative. The minimum inhibitory concentration (MIC) values for the isolated M. hominis, measured by the microdilution method, were < 0.12 μg/ml and < 0.06 μg/ml for clindamycin and ciprofloxacin, respectively. Each serum and CSF sample was obtained almost 2 hours before the administration of ciprofloxacin and 4 hours before the administration of clindamycin. At all times, the concentrations of the 2 drugs in the serum and the CSF samples were above the MIC of the isolated M. hominis. The concentrations of these drugs in the CSF samples are shown in Fig. 3. The patient was discharged, fully recovered, on her 50th day of hospitalization. At 1 year after the VPS reinsertion, the patient was healthy, without sequelae, and with no evidence of any infection recurrence.

Image showing pinpoint colonies grown from a pus sample on a chocolate agar plate. Pinpoint colonies grew from pus samples on both blood agar and chocolate agar plates. At a later date, the same type of colonies also grew from CSF samples. Figure is available in color online only.

Ciprofloxacin (CPFX) and clindamycin (CLDM) concentrations in the patient's CSF. The determined concentrations of the 2 drugs remained consistently above the MIC of the isolated M. hominis at all times (dashed line shows the MIC of clindamycin and dashed-and-dotted line shows the MIC of ciprofloxacin). In addition, the concentration of each drug in the serum remained consistently above the MIC at all times. Figure is available in color online only.

Discussion

Infections are known to occur following VPS insertion in approximately 3%–20% of patients (about 3% in our institute in the past 3 years), and these infections can occur either with or without CNS involvement.^14,15 M. hominis has been rarely reported as a causative agent for VPS infection in the past and the appropriate treatment for such infection remains unclear. The only previously reported symptomatic case of VPS infection caused by M. hominis was an 11-year-old girl with hydrocephalus.¹⁰ In this case, the shunt became infected just 3 weeks after insertion and M. hominis was cultured from the patient's CSF, which also showed pleocytosis. This patient was first treated with vancomycin plus erythromycin. Although in this case the patient's symptoms and CSF findings were initially relieved, the infection recurred following treatment. The patient was then administered methacycline and the VPS was removed. Unfortunately, the patient died during this therapy.

Some cases have also been reported of M. hominis infection with CNS involvement following neurosurgical procedures. Whitson et al.²¹ reported 11 postoperative cases of M. hominis CNS infections following neurosurgical intervention. All patients except an 11-year-old girl with scoliosis were adults. None of these patients underwent a VPS insertion procedure. Although inappropriate therapies were administered and washout surgeries were required for some of these patients, most of the patients were treated with tetracyclines and/or fluoroquinolones and after diagnosis they had good outcomes.

In our case, the VPS infection with CNS involvement was successfully treated with intravenous clindamycin plus ciprofloxacin therapy in addition to VPS externalization, EVD revision, and subsequent VPS replacement. Although M. hominis is not generally susceptible to antibiotics used for pediatric infections, previously reported cases have demonstrated that the optimal antibiotic agents against M. hominis include chloramphenicol, clindamycin, quinolones, and tetracyclines.⁴ Although tetracyclines can achieve good CSF penetration, their undesirable adverse effects on teeth and bones in children have led to a general hesitation when it comes to their application. Chloramphenicol has also been associated with the risk of irreversible aplastic anemia. For these reasons, quinolones and clindamycin appeared to be good options and we thus treated our current patient with both of these drugs. Ciprofloxacin, which we administered to the patient as a quinolone agent, has also been reported to be associated with the risk of adverse musculoskeletal events, but these events are reversible.¹

In our therapy of the patient with clindamycin and ciprofloxacin, monitoring of the concentrations of the 2 drugs in the patient's serum and CSF and repeated PCR of M. hominis in her CSF were also performed, for the following reasons: 1) parenteral moxifloxacin, which has been shown to be more effective against M. hominis infections among the quinolones,² is not available in Japan; 2) clindamycin is known for its poor CSF penetration; and 3) there has been no report of a successfully treated case with VPS infection caused by M. hominis in the past, so no data were available to support the use of these drugs to treat this type of infection. We also used the microdilution method to determine the MIC values of the 2 drugs for the M. hominis isolated from cultures and verified that the concentrations of both drugs consistently remained above these MIC values. The results suggested that the ciprofloxacin and clindamycin were effective in our case.

The patient required 58 days of hospitalization, which seemed to be a long period of inpatient care for a VPS infection. According to previously reported cases, treatment durations of M. hominis extragenital infections vary widely. Because there was no previous report of a successfully treated case with VPS infection, we decided to pursue a 6-week course of parenteral antibiotic administration. The duration of therapy was determined from the 1st day of PCR-negative results of the patient's CSF sample. This 6-week period is one of the longest reported treatment durations; however, because of the lack of data from similar cases, it was difficult to determine the appropriate length of treatment for this infection.

In some reported cases, particularly in neonates, patients have presented with no symptomatic conditions and no treatment has been required despite the isolation of M. hominis from their CSF.^16–19 However, in many cases patients have died or suffered from permanent neurological sequelae. Because our experience of M. hominis infections in children is very limited, we urgently need to conduct further research to investigate the significance and natural history of these infections. In our case, as well as in previously reported cases, M. hominis infection presented an invasive course but was controlled soon after treatment was changed to an effective antibiotic regimen.

Conclusions

As our case highlights, when beta-lactam antibiotics are ineffective against VPS infections, and when Gram staining of clinical specimens is unable to detect the causative pathogens, then it is important to consider M. hominis as the atypical pathogen. In addition, our combined treatment with ciprofloxacin plus clindamycin was effective against the VPS infection with little concern for possible adverse effects of these drugs in a pediatric patient.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Sato. Acquisition of data: Sato, Kubota, Katsuyama, Suzuki, Miyairi. Analysis and interpretation of data: Sato, Kubota, Katsuyama. Drafting the article: Sato. Critically revising the article: Kubota, Katsuyama, Suzuki, Minami, Kasai. Reviewed submitted version of manuscript: Miyairi, Minami, Kasai. Approved the final version of the manuscript on behalf of all authors: Sato. Administrative/technical/material support: Kubota, Katsuyama.