Introduction

Patients admitted into intensive care units (ICUs) are at a great risk for acquiring nosocomial infections due to their critical conditions and weakened immune system [1]. Ventilator-associated pneumonia (VAP) is the most frequent nosocomial infection in the ICU and occurs in roughly 25% of patients requiring mechanical ventilation (MV) for more than 48 h [2, 3]. Mycoplasma hominis is a commensal inhabitant of the lower urogenital tract in many sexually active adult men and women [4]. However, it may also cause localized extragenital disease. Host predisposing factors such as immunosuppression, malignancy, trauma, and complications related to urogenital manipulations or surgery are considered to be associated with extragenital mycoplasma infections [5]. The recognition of infections by Mycoplasma species in critically ill patients is important because β-lactam and aminoglycoside antibiotics, which are usually used in empirical antibiotic therapy, are not effective against these organisms. Pneumonia caused by M. hominis is very rare, but since the true incidence of infections caused by this pathogen is unknown, its real importance may be underestimated. Recognition of the morphological characteristics of mycoplasmas on bacteriological culture media may increase the recovery of these organisms in atypical clinical settings [6]. Here we describe the clinical and microbiological features of seven isolates of M. hominis in critically ill patients who acquired pneumonia during their hospitalization to the ICU. The case of one of these patients has been previously reported [7].

Patients and methods

Patients

Between April 2000 and December 2004 seven patients were found to have respiratory infections by M. hominis in three ICUs at the University Hospital Marqués de Valdecilla. All of these isolates were recovered from adult men who presented clinical signs of pneumonia as well as images of pulmonary infiltrates on chest radiography, temperature of 38 °C higher or of 36.5 °C or lower, peripheral leukocyte count greater than 11,000 or less than 5,000 cells/μl, and the presence of purulent secretions.

Specimen processing and isolates

Pleural fluid and bronchoalveolar lavage (BAL) specimens were processed by standardized methods [8] and were tested for Gram-stain and aerobic-anaerobic culture. The incubation time extended for up to 72 h. Owing to the morphology and size of the colonies and the fact that Gram-stained smears showed no micro-organisms, the presence of Mycoplasma spp. was suspected, and colonies were therefore subcultured into UMM lyo broth with Mh supplement (Mycofast Evolution 2 International Microbio Signes, France). Definitive identification was determined by automated DNA sequencing of 16S rRNA gene and sequence comparison from sequences available from the GenBank Nucleotide Sequence Database [9]. In addition, all samples were tested for the presence of other respiratory pathogens, such as Legionella, fungi, mycobacteria, and viruses.

Antibiotic susceptibility testing

A broth microdilution method for determination of the minimal inhibitory concentration was performed using Sensititre panels (TREK Diagnostic Systems. Cleveland, Ohio, USA) [10]. Erythromycin resistance was confirmed by the Etest method (Biodisk, Solna, Sweden) [11].

Results

Demographic information, clinical and microbiological findings, and outcomes of the seven cases are summarized in Table 1. The patients' median age was 46 years (range 25–65). No patient had a respiratory tract infection when admitted to the ICU and all had been treated with empirical antibiotic therapy prior to diagnosis of pneumonia due to prophylaxis after lung transplant, multiple trauma or sepsis. Four patients required MV for postoperatory situation, polytraumatism, sepsis, and coma, and three patients were weaned off MV prior to pneumonia. Symptoms of pneumonia started 96 h or later after hospitalization in the ICU. In all patients BAL were obtained when the start of pneumonia symptoms was observed, after which empirical antibiotic therapy was administered. In addition, pleural fluids were obtained on patient numbers 5 and 6 when empirical antibiotic therapy had been started. M. hominis was the only pathogen isolated from the site of infections in four cases and was associated with a copathogen in the remaining three (Table 1). M. hominis diagnosis took an average of 8.4 days (range 6–17 days) after the collection of respiratory samples. M. hominis was not recovered in blood cultures performed at the beginning of pneumonia symptoms. Results for Legionella, fungi, mycobacteria, and virus cultures all were negative. Urines were cultured specifically only for urogenital mycoplamas in three patients and M. hominis along with Ureaplasma urealyticum was isolated from two of them. The sensitivity of M. hominis isolates to the different antibiotics tested is given in Table 2. Once M. hominis was diagnosed, empirical treatment was changed to an effective antibiotic, except in patient numbers 4 and 7, in whom M. hominis pneumonia was diagnosed postmortem. The clinical course was torpid in all of them, although in patient numbers 2, 3, and 5 the specific treatment for M. hominis with quinolones for 2 weeks was associated with a good clinical response. In patient number 1, M. hominis diagnosis and treatment change to quinolone was made 2 days before its death. Autopsies confirmed the causes of death (Table 1).

Table 1 Selected demographic data, clinical and microbiological findings, and outcome in seven male patients with Mycoplasma hominis infections during ICU hospitalization. (ICU intensive care unit, BAL bronchoalveolar lavage, PDA parenteral drug addict, COPD chronic obstructive pulmonary disease, piper/tazob piperaciline-tazobactam, MOF multiple organ failure)
Full size table
Table 2 Antibiotics tested for Mycoplasma hominis isolates. The Clinical Laboratory Standard Institute has not yet published official breakpoints for Mycoplasma; for this reason resistance and susceptibility were defined according to those established for fastidious organisms (MIC minimal inhibitory concentration, R resistance, S susceptible)
Full size table

Discussion

M. hominis is generally considered an opportunistic agent that causes invasive infections in susceptible populations [12, 13]. Accordingly, critically ill patients in ICUs could be more receptive to be infected because of their decreased immunity and frequent use of invasive procedures. VAP, the predominant infection in ICUs, is most commonly due to potentially multiresistant bacteria [14], and consequently prevention and control strategies have focused on methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococcus, Pseudomona aeruginosa, Acinetobacter spp., and extended-spectrum β-lactamase-producing Gram-negative bacilli, among others. Outcome of patients with ICU-acquired pneumonia depends on the adequacy of early empirical antibiotic therapy [3, 15] which is often based on timing of its occurrence in relation to the onset of MV. M. hominis is intrinsically resistant to β-lactams, macrolides, aminoglycosides, and glycopeptides which are commonly used as empirical therapy in ICU infections. The real incidence of pneumonia caused by M. hominis in ICU is unknown and seems to be underestimated. In our center 512 samples of BAL or pleural fluid were tested by conventional bacteriological culture between 2000 and 2004. The incidence of M. hominis pneumonia was estimated to be 1.3% among samples tested.

In patients admitted to the ICU the presence of an invasive device as orotracheal tube or urethral catheters increases the risk of infection by favoring the access of micro-organisms into a normally sterile space. The origin of the M. hominis infection in these patients is unknown. A possible explanation is microaspiration of oropharyngeal contents, but we consider that our isolates of M. hominis did not represent oropharyngeal contamination because the proportion of squamous epithelial cells, as indicator of oropharyngeal contamination, was less than 1% in all episodes. Morever, all specimens showed heavy growth of M. hominis in cultures. We also excluded other causes to explain clinical signs of pneumonia, such as a lesional edema complicating a thoracic trauma, a reperfusion injury after lung transplantation, and a lung transplant rejection.

A possible alternative source of infection is mycoplasmaremia from the colonized urogenital tract and secondary to urethral catheterization. Unfortunately, we did not obtain urine or urethral discharge samples from all of our patients in order to investigate the presence of urogenital mycoplasmas. M. hominis was not recovered in blood cultures since mycoplasmas cannot be detected by routine blood culture systems, probably due to the mycoplasmastatic effects of the sodium polyanethol sulfonate, the anticoagulant widely used in liquid blood culture media [16, 17]. Suspicion of Mycoplasma infection must be heightened particularly in critically ill patients for whom there are findings of purulence in the face of negative results from cultures and smears. Nevertheless, VAP is polymicrobial in nearly 40% of critically patients [18]. In our series the concomitant P. aeruginosa and/or A. baummanii isolation diminished the importance of the pathogenic role of M. hominis.

One factor that is potentially associated with a more severe outcome for patients with polymicrobial infections would be the inappropriateness of initial antibiotic therapy, with some bacterial species escaping the spectrum of initial empirical therapy. Antimicrobials agents that interfere with protein synthesis, such as tetracyclin and clindamycin, and those that inhibit DNA replication, such as quinolones, are considered effective against M. hominis. In three patients the specific treatment for pneumonia with quinolones was associated with a good clinical response. Linezolid has also proved to have some activity [19], but it was not assayed in this work. There is evidence suggesting that excessive mortality due to the administration of inappropriate antibiotics is not reduced by correction of regimens when culture results are available later [9, 20]. Although it is difficult to believe that M. hominis pneumonia was the cause of death in the four patients, these infections may have contributed to increased suffering and costs for patients in the form of prolonged stays.

Although M. hominis can grow on conventional microbiological media, more specific laboratory methods and a prolonged period of incubation should be used in case this organism is suspected. Although methods are available for molecular detection of specific agents involved in determined respiratory infections (e. g., multiplex polymerase chain reaction for detection of atypical pneumonia agents), similar assays for the empirical detection of unexpected pathogens responsible for cases of undetermined pneumonia are currently unavailable. Recognition of the potential for M. hominis to cause pneumonia should prompt request for special microbiological studies, especially in critically ill patients who present with pulmonary infiltrates and negative cultures for respiratory pathogens, despite the presence of neutrophils on Gram's stains of respiratory tissue or fluid specimens. In summary, it is prudent to consider M. hominis as a potential opportunistic pathogen in acquired pneumonias in ICUs, while further research is needed to increase our understanding in the epidemiology and the outcome of this unusual infection.