Background

Parvimonas micra (P. micra), also known as Peptostreptococcus micros and Micromonas micra, can be commonly detected on the surface of human skin and as a part of the dental and gastrointestinal flora. We have found multiple case reports on P. micra-related periodontitis, pylephlebitis [1], iliopsoas abscess [2], abdominal abscess [3], and arthritis [4]. However, we have not found any reports on P. micra-related pneumonia. Because the lung is an oxygen-containing organ, it is an unsuitable environment for fastidious Parvimonas micra.

Due to difficulty in culturing the bacteria, P. micra infection can be missed and thus its treatment delayed, possibly resulting in a worse prognosis [4]. In this case, we used next-generation sequencing (NGS) to detect P. micra-related pneumonia.

Case presentation

A 35-year-old pregnant woman was admitted to our emergency room 4 weeks prior to her due date because of a sudden headache and unconsciousness. Computed tomography (CT) showed a massive cerebral haemorrhage in the right hemisphere, and she was immediately taken to the operation room for an emergency caesarean section. After the foetus was delivered, she underwent a craniotomy to remove the cerebral haematoma. After 8 h of surgery, the patient was taken to the ICU for further treatment.

The patient developed dilated pupils without a pupillary response. The CT scan showed mild pneumonia (Fig. 1a and b) on day 1. Moxifloxacin was started to treat her pneumonia and to prevent surgical site infections. Laboratory examinations on admission revealed a leukocyte count of 10*10^9/ml and a C-reactive protein level of 21.9 mg/L. As the pneumonia progressed, the patient developed febrile fever with a temperature between 38.5–39.5 °C and developed more grossly purulent tracheobronchial secretions, and laboratory studies showed an increased leukocyte count. Meanwhile, no pathogen was found as multiple sputum cultures, blood cultures and IgM for Chlamydia and Mycoplasma all came back negative, which ruled out common hospital acquired pneumonia caused by Streptococcus pneumoniae, Klebsiella pneumoniae or Staphylococcus aureus. Despite starting moxifloxacin at admission and meropenem and vancomycin on day 3 after admission, her pneumonia worsened, and she soon developed hypoxemia (oxygenation index 55–85 mmHg). The CT performed on day 12 showed more exudation and atelectasis than previously (Fig. 1c and d).

Fig. 1
figure 1

CT findings of lung: a and b The computed tomography (CT) showed mild exudation in the lower lobe of the lung which revealed a pneumonia at admission. c and d The CT on day 12 showed the involvement of the lung had progressed. e and f The CT scan on day 24 showed clearer lung field and less infiltration and atelectasis, and the pneumonia was improved

Full size image

Since the bacteriological culturing protocol failed to demonstrate a responsible pathogen and our current antibiotic therapy failed to control the pneumonia, we came to realize that the pathogen could be an uncommon pathogen for severe pneumonia. We sent a sample of bronchoalveolar lavage fluid for pathogen detection by next-generation sequencing (NGS) at BGI-Shenzhen on day 13 after admission. The NGS test was performed on the BGISEQ-500 platform [5]. Within 24 h, the NGS results identified 252 DNA sequence reads (out of 18,188,496) and 621 RNA sequence reads (out of 38,214,718) corresponding to P. micra. After we eliminated all the sequence reads of the human host and the unclassified reads, P. micra reads accounted for 83.17% (DNA) and 60.53% (RNA) of the total microbial reads (Fig. 2). The second pathogen identified by NGS was Stenotrophomonas maltophilia, which could be a co-pathogen. Thus, the NGS results suggested that P. micra was the main pathogen causing pneumonia. On day 14, we adjusted the antibiotic therapy from meropenem and vancomycin to ornidazole for P. micra and sulfamethoxazole for Stenotrophomonas maltophilia according to our local antibiotic sensitivity pattern.

Fig. 2
figure 2

NGS findings on day 13: a NGS detected the total DNA reads, 97.8% of the reads were from the human host, 2.15% of the reads were too small to be classified, only 0.05% of the reads were bacterial. b Among the bacterial DNA reads, 83.17% of the reads indicated P. micra infection. c NGS detected the total RNA reads, 98.3% of the reads were from the human host, 1.58% of the reads were unclassified, 0.05% reads were bacterial. d 60.53% of the bacterial RNA reads indicated P. micra infection

Full size image

After starting ornidazole, a prone position was given and daily bronchoscopic sputum suction, the hypoxemia improved. A CT scan performed on day 24 found less infiltration (Fig. 1e and f) and fewer lesions of atelectasis. Bronchoscopy showed less sputum and less bronchial oedema.

Unfortunately, due to the patient’s massive cerebral haemorrhage, she was diagnosed with brain death based on electroencephalogram (EEG) and her clinical status.

Discussion and conclusions

P. micra is a part of the normal flora of the oral cavity, gastrointestinal tract, genitourinary tract and skin. P. micra is not a dominant bacterium in an aerobic environment, and there have been multiple case reports on P. micra-related abdominal infection [1,2,3, 6, 7], spondylodiscitis [8, 9], intracranial infection [2, 10, 11] and bloodstream infection [12,13,14,15], but we did not find any case reports describing severe pneumonia caused by P. micra. On the other hand, we noticed that P. micra could cause deep infection in the elderly patients or patients who have undergone surgery and cancer therapy exist [11, 12, 14, 16], indicating that immune insufficiency and postoperative stress could be risk factors for this infection. Our patient had no history of immune insufficiency or cancer, so we suspected that stress after surgery was the main risk factor for this atypical pneumonia.

In this case, the patient was already in a coma before admission to our ER; hence, aspiration pneumonia caused by bacteria in the oral cavity and gastrointestinal tract was taken into consideration. We placed the patient on a ventilator because she had no spontaneous breathing or cough reflex; her lung was filled with thick sputum that enlarged the volume of respiratory dead space, creating a partial anaerobic environment for P. micra. After we identified the pathogen with NGS, we used ornidazole to eliminate P. micra and prone position ventilation along with bronchoscopic sputum suction to reduce the volume of respiratory dead space. Once P. micra was eliminated and the partial anaerobic environment was improved, the infection was quickly placed under control.

Another bacterium identified by NGS was Stenotrophomonas maltophilia, which is a common cause of ICU-acquired pneumonia, especially among immunocompetent patients. After a series of consultations with the aetiologist, we thought Stenotrophomonas maltophilia might not be the main cause of this pneumonia because the pneumonia still progressed despite the application of a strong antibacterial therapy including moxifloxacin, meropenem and vancomycin. However, Stenotrophomonas maltophilia could still contribute to a co-pathogen according to the NGS results.

There were several limitations in this case report. We did not find any bacteria or fungi in blood or sputum culture, and we also did not find P. micra, this could be due to the lack of anaerobic bacteria culture facility in our centre. Second, we did not submit bronchoalveolar lavage fluid for next-generation sequencing (NGS) after ornidazole therapy to confirm whether the P. micra was eradicated due to the expensive cost of this test. The diagnosis of P. micra-related severe pneumonia was determined by the NGS results, the lack of evidence of other common pathogens, and the curative effect of ornidazole therapy.

In conclusion, identifying the pathogen is the key factor for treating pneumonia. When bacteriological culturing fails, NGS may play an important role in quick and precise diagnosis.