Introduction

Phytobacter diazotrophicus (P. diazotrophicus), a gram-negative bacterium, is an opportunistic pathogen [1]. In 2008, P. diazotrophicus was first described as an endophytic enterobacterial species isolated from wild rice in China [2]. In Brazil, P. diazotrophicus-contaminated total parenteral nutrition has led to infection outbreaks in neonatal intensive care units [3]. Although P. diazotrophicus has only been recently described in human samples [4, 5], WGS revealed that it has been frequently misidentified as Pantoea, Metakosakonia or Kluyvera since the 1970s [3, 5]. Pantoea spp. (later corrected to Phytobacter spp [5]. ) outbreaks have resulted in high neonatal mortality [6]. To date, P. diazotrophicus has not been isolated from blood in China.

Galactosemia is an autosomal recessive inherited metabolic disorder of galactose metabolism. GALAC1 is caused by mutations in the GALT gene mapped to chromosome 9p13; it is the most frequent form of galactosemia, with a prevalence of approximately 1 in 60,000 live births in the general population [7].

Here, we report a rare case of neonatal sepsis caused by P. diazotrophicus combined with GALAC1.

Case presentation

A 27-day-old female infant who presented with fever half a day ago and high bilirubin levels was admitted to hospital. She was delivered vaginally at 37+ 3 weeks, with no history of resuscitation for asphyxia, birth weight of 2680 g, and Apgar score of 10-10-10. After birth, she was admitted to a local county hospital for neonatal hyperbilirubinemia, disseminated intravascular coagulation, and neonatal ABO haemolytic jaundice. During the 20 days of initial hospitalization, she received treatments such as infusion of human immunoglobulin, human albumin, fresh frozen plasma, low molecular weight heparin sodium, cryoprecipitation coagulation factors, and leukoreduced red blood cells. At 4 days old, tandem mass spectrometry was performed for neonatal genetic metabolic disease screening and revealed increased tyrosine, phenylalanine, and citrulline levels, which were ignored during the initial hospitalization.

Upon physical examination, her body temperature, blood pressure, pulse rate, and respiratory rate were 37.5°C, 88/44 (55) mmHg, 150 beats/min, and 52 breaths/min, respectively. The infant exhibited yellowish skin, muscle weakness, limb convulsions, an incomplete hugging reflex, and a weak sucking reflex. The laboratory results were as follows: low blood glucose (1.3 mmol/L), high neonatal total bilirubin (340 umol/L), high direct bilirubin (229.9 umol/L), high total bile acid (201.3 umol/L), marginally elevated alanine aminotransferase (53 U/L), marginally elevated aspartate aminotransferase (73 U/L), elevated C-reactive protein (18.43 mg/L; normal range:≤1.6 mg/L), elevated interleukin-6 (109 pg/mL; normal range:18–26 pg/mL), elevated procalcitonin test (11.74 ng/mL; normal range: ≤0.5 ng/mL), prolonged prothrombin time (> 120 sec), decreased fibrinogen (< 0.5 g/L), prolonged activated partial thromboplastin time (> 180 sec), and decreased haemoglobin (87 g/L). The direct antiglobulin test (Coombs’) results were positive for anti-human IGG and negative for anti-human C3. Urinalysis revealed +++ urinary bilirubin. Cerebrospinal fluid analysis was normal. Anteroposterior chest radiograph revealed a left clavicle fracture. Routine colour ultrasound of the infant’s belly revealed a slightly enlarged liver, thickened gleasonian sheath, enhanced parenchymal echo of both kidneys, and large amounts of ascites. Cranial colour ultrasonography revealed cerebral oedema. Using a video electroencephalogram, focal epileptic electrical seizures were detected, with numerous left occipital area waves observed between ictal statuses.

On the third day of hospital admission, two blood cultures from two different body parts revealed gram-negative rods; therefore, neonatal sepsis was considered. Anti-infective treatment with piperacillin/tazobactam was administered before culture report. VITEK-2 GN card (bioMérieux) displayed acceptable identification of Kluyvera intermedius (93%). The MALDI Biotyper-Bruker Microfex showed that Phytobacter ursingii, which has a spectral score of 1.9 (low confidence), was the closest match, and the strain was subjected to WGS. Multilocus sequence analysis (MLSA) concatenated with five housekee** genes (dnaJ, mdh, pyrC, recA, and rpoD), which is an effective method for identifying bacterial species, was used to perform phylogenetic analyses, using the MEGA11 software. The closest affinity of Pd1 was with P. diazotrophicus ENNIH2 and ENNIH3 (Fig. 1). To further identify the strain, genomic relatedness among different species was determined using web-based DNA-DNA hybridisation, including in silico DNA-DNA hybridisation (dDDH) and average nucleotide identity (ANI). The dDDH and ANI values for strains Pd1, P. diazotrophicus ENNIH2, P. diazotrophicus ENNIH3, P. diazotrophicus A-F18, P. diazotrophicus TA9759, and P. diazotrophicus TA9832 ranged from 78.5 to 99.9% and 97.5–99.9%, respectively (Table 1 and Table 2; above the 70 and 95% thresholds), which indicated that strain Pd1 was P. diazotrophicus. Finally, in this study, we determined that the causative neonatal sepsis pathogen was P. diazotrophicus and not K. intermedius, as reported by the laboratory.

Table 1 Digital DDH of strains using genome-to-genome distance calculator
Full size table
Table 2 Mean ANI values for all available genomes of strains
Full size table
Fig. 1
figure 1

Phylogenetic analysis between the strains. The strains genomes available on the NCBI database and Pd1 isolated in this study

Full size image

Based on the combination of the tandem mass spectrometry results and clinical features, inherited metabolic diseases need to be considered. Therefore, WES was performed for the patient and her parents to identify the mutations in the patient. Two pathogenic/likely pathogenic compound heterozygous mutations were identified: a nonsense mutation NM_000155: c.610 C > T:p. Arg204Ter in exon7 (pathogenic) and a missense mutation NM_000155: c.1034 C > A:p. Ala345Asp in exon10 (likely pathogenic) of GALT gene, which explains the patient’s clinical phenotype. The patient’s parents were heterozygous carriers of both mutations. Compound heterozygous mutations in GALT were consistent with the autosomal recessive inheritance pattern of GALAC1. Therefore, based on confirmed GALAC1, the newborn was switched to lactose-free milk powder.

The antimicrobial susceptibility of the isolated strain was determined using the VITEK 2 bacterial antibiotic sensitivity analysis system (bioMérieux). The results were interpreted based on breakpoints recommended by the Clinical and Laboratory Standards Institute M100-S26 for Enterobacteriaceae. Based on the antimicrobial susceptibility results (Table 3), the antimicrobial therapy was changed to meropenem. Meropenem was administered for 10 days to fight the infection until the infection index returned to the normal range. The 22-day comprehensive treatments also included the multiple transfusion of red blood cells, filtered fresh frozen plasma, cryoprecipitate coagulation factors, fibrinogen, albumin gamma globulin, and low molecular weight heparin, dehydration and diuresis to reduce intracranial pressure, ursodeoxycholic acid to protect the liver and gallbladder, ascites drainage, intravenous nutrition, and phenobarbital sedation maintenance.

Table 3 Antibiotic resistance result for strain Pd1
Full size table

Discussion and conclusions

In this study, we found that phenotypic identification methods could not identify P. diazotrophicus, which is consistent with a previous report [4]. The in-house SuperSpectrum for MALDI-TOF MS can identify Phytobacter spp [5], however, we repeatedly failed to successfully identify P. diazotrophicus, possibly because we used instruments from different manufacturers. WGS is the most reliable approach for identifying suspected isolates at the species level using ANI, dDDH, or core genome phylogeny [5].Through ANI, dDDH, and MLSA phylogeny, P. diazotrophicus was successfully identified.

The OMIM database (No.230400) indicated a high incidence of E. coli sepsis in untreated neonates GALAC1 [8], suggesting that galactosemia may be an important factor involved in sepsis. It is worth noting that the isolate in this study was not E. coli but P. diazotrophicus, which facilitated the identification of the GALAC1 clinical features. From this case, we found that GALAC1 disease required multiple infusion treatments for a long time in the neonatal ward. This may be an important factor in the occurrence of P. diazotrophicus bloodstream infection for neonate. A diet that minimises galactose intake is the cornerstone of treatment for GALAC1 [7]. In the present study, a lactose-free milk powder diet improved the main symptoms of GALAC1 in the infant.

In recent years, reports of multidrug-resistant P. diazotrophicus have been increasing. Concerningly, the strains primarily carried carbapenem resistance genes, blaNDM−1 or blaKPC, on plasmids resistant to most β-lactam antibiotics, which can easily cause horizontal transmission [1, 9]. Fortunately, no carbapenem resistance genes were found in this study, and meropenem treatment ultimately controlled the infection.

In conclusion, we successfully identified P. diazotrophicus in neonates with sepsis in China and isolated P. diazotrophicus from patients with galactosemia. Galactosemia may be an important cause of neonatal sepsis. These findings further expands the current understanding of the clinical characteristics of GALAC1.