An extremely premature neonate born at gestational age 24 + 2 weeks and days with birth weight 765 g, admitted to NICU at Stavanger University Hospital, Norway November 2020 (Figure 1). The mother was GBS-negative but nonetheless received intrapartum benzylpenicillin at a dose of 1.2 g six times daily from the time of admission, and this regimen was continued for 48 h. The neonate was born after vaginal delivery, and the APGAR score was 8–9–10. Although breathing spontaneously at 1 min, the neonate was intubated shortly after birth for administration of surfactant. Venous and arterial umbilical catheters and a nasogastric tube were inserted. First line empirical antibiotics benzylpenicillin and gentamicin were administered but discontinued after 48 h. In line with national guidelines, the neonate received antifungal prophylaxis (mycostatin). The neonate was extubated day 2 to continuous positive airway pressure (CPAP). Echocardiography revealed a small patent foramen ovale, otherwise the cardiological examination was normal.

On day 12, the neonate presented with increasing apnoea episodes, and small sores with secretion were observed in both nostrils. The level of C-reactive protein (CRP) was <1 mg/L, hemoglobin 13.2 g/dL, pH 7.26 and pCO₂ 6.4 kPa. Chest X-ray was normal. Blood cultures were negative. Benzylpenicillin and gentamicin were administered on suspicion of sepsis but discontinued after 48 h. CPAP was discontinued in favor of low-flow nasal cannula day 12.

On day 27, the neonate again suffered from increasing apnoea episodes, in addition to intolerance for enteral feeds and increased abdominal distention. CRP was slightly elevated at 16 mg/L. Benzylpenicillin and gentamicin were for the third time administered for 2 days before being discontinued. PCR of nasopharynx secretion (FilmArray, BioFire® Respiratory Panel 2.1 plus) was negative. Blood cultures were negative. However, swabs from nostrils and nasal sores showed growth of S. lugdunensis and S. aureus 1 day later. The neonate improved clinically without need for further antibiotic treatment. The neonate was discharged from the hospital 12 weeks after birth.

Cultivation of swabs from the nasal sores revealed growth of S. lugdunensis (MALDI-TOF score 2.26) and S. aureus. AST revealed a S. aureus with no phenotypic resistance to tested antibiotics. AST of S. lugdunensis using cefoxitin disk diffusion yielded a zone diameter of 25 mm. With the cefoxitin breakpoint at the time of testing, this would be interpreted as susceptible. However, due to the locally introduced ATU, the strain was selected for mecA/C-PCR, which confirmed the presence of mecA.

Although the CHROMAgar™ MRSA plates are not specifically validated for the detection of MRSL, the MRSL isolates from screening samples grew on the plate in this case, and the additional blood agar supplemented with antibiotics was not necessary for detection. S. lugdunensis grew with slightly yellow, shiny colonies with β-hemolysis on blood agar, and with similar colonies on CHROMAgar™ MRSA (Figure 2).

Further susceptibility-testing of the S. lugdunensis strain revealed a cefoxitin minimum inhibitory concentration (MIC) of 32 mg/L (resistant > 4 mg/L). By disc diffusion, the strain was furthermore resistant to gentamicin and tetracycline, susceptible with increased exposure (I) to ciprofloxacin, and susceptible to the remaining antibiotics tested (Table 1).

After the detection of MRSL in the clinical sample, the hospital infection control team was consulted. The patient and her parents were isolated in a separate room with contact isolation and were screened for MRSL colonization by the same local protocol as for MRSA. Screening of the neonate revealed colonization of MRSL in the nose, throat and perineum, while her parents were screening negative. However, 2 months later, the mother was diagnosed with mastitis by her general practitioner, and MRSL was subsequently detected in a breast milk sample. At 6 months follow-up, the girl and her parents were MRSL negative, and no further screening was performed.

Detection of MRSL in the index patient prompted screening of other infants in the NICU, their parents, and clinical personnel using CHROMAgar™ MRSA plates followed by MALDI-TOF identification. As reported previously, this revealed that four patients and three staff members were colonized. Furthermore, two patients developed sepsis, one of whom had MRSL in blood culture (12).

The MRSL strain isolated from the premature neonate was whole genome sequenced, yielding a 2,594,971 bp chromosome and a 45,744 bp plasmid named pMRSL-1 (Figure 3). The genome has a total GC content of 33.7% and encodes 2,752 protein-coding genes and 5 rRNA operons. In-silico typing revealed that the strain belonged to ST3 and had a chromosomal class IV (2B) SCCmec with mecA and chromosomal blaZ. Other genes conferring antibiotic resistance were plasmid-located, and included aadD, tet(K), lnuA and a truncated, yet functional aac(6′)-aph(2″)-gene. The plasmid was potentially mobilizable (class MOB_V) and belonged to plasmid taxonomic unit (PTU) Bac21. It had a grade II host-range (13), which suggests a narrow host distribution, and closely related plasmids (mash distance ≥0.99 and p-value <0.01) were primarily found in other staphylococci, mainly S. aureus (43.8%) and S. epidermidis (15.6%). Although similar multidrug-resistance plasmids have been characterized in S. lugdunensis previously (14), the best hit using nBLAST covered only 64% of the plasmid.

There were detected 34 virulence-associated genes in the MRSL strain, including type 8 capsular polysaccharide genes, a type VII secretion system and iron acquisition system, as well as a potential sphingomyelinase-c toxin (Supplementary Table S1). Comparative analysis with the functional lugdunin operon of S. lugdunensis N920143 (15) demonstrated that lugA, lugB, and lugC in the MRSL strain each harbor multiple frameshift mutations that introduce internal stop codons. These disruptions indicate that the lugdunin biosynthetic operon is degenerated and unlikely to encode a functional antimicrobial product.

Clinical data, including gestational age, birth weight, treatment and complications during NICU stay, was collected retrospectively from the patient’s electronic medical journal.

Screening samples of the patient and parents from nose, throat and perineum were obtained and placed in Amies transport medium (ESwab®; Copan). The samples were cultivated on chromogenic agar supplemented with methicillin (CHROMAgar™ MRSA; CHROMAgar) routinely used for MRSA-screening. MRSA displays pink to mauve colonies on this agar, while other bacteria may have blue, white or slightly yellow colonies. Screening samples were simultaneously cultivated on sheep blood agar supplemented with colistin and aztreonam for selection of Gram-positive bacteria. Agar plates were incubated at 35 °C for 48 h, and growth evaluated every 24 h. Suspected S. lugdunensis colonies were identified by MALDI-TOF (Bruker Daltonics). Confirmed S. lugdunensis strains were subjected to PCR (GeneXpert® MRSA NxG; Cepheid) for mecA/C detection.

Antimicrobial susceptibility testing (AST) was performed according to the European committee on antimicrobial susceptibility testing (EUCAST) (26) on Mueller Hinton agar (OXOID). Screening for methicillin-resistance was performed with cefoxitin disk diffusion (30 μg, OXOID). Further susceptibility-testing was performed with agar gradient diffusion methodology (MIC Test Strip, Liofilchem).

The EUCAST cefoxitin breakpoint for S. aureus and CoNS (other than S. epidermidis) was 22 mm (sensitive ≥ 22 mm, resistant < 22 mm) and the epidemiologic cut-off value (ECOFF) 22 mm at the time of testing (20). For CoNS not identified to species level, a zone diameter of ≥25 mm was considered susceptible and <22 mm resistant. In our laboratory, a defined area of technical uncertainty (ATU) was applied during antimicrobial susceptibility testing (AST) of S. aureus, S. lugdunensis, and S. argenteus against cefoxitin (30 μg). Isolates exhibiting inhibition zones between 22 and 26 mm were subjected to mecA/C PCR analysis when deemed clinically relevant. The adjusted ATU was introduced after repeated observations of unusually large cefoxitin inhibition zones in our quality-control strains, raising concerns that the standard criteria might underestimate resistance. This modification was therefore implemented to reduce the risk of false-susceptible results in routine screening.

Whole genome sequencing (WGS) was performed at the Norwegian MRSA reference laboratory, St. Olav’s hospital, Trondheim University hospital. Briefly, isolation of DNA was performed with an EZ1 Advanced XL System (Qiagen), while WGS was performed on a MiSeq system, using the Nextera XT DNA sample preparation kit and MiSeq® Reagent Kit v3 (600 cycle; Illumina). The strain was also sequenced using the rapid sequencing kit (SQK-RAD004) with a flongle flow cell (FLO-FLG001) on a MinION instrument (Oxford Nanopore Technologies). The nanopore data were assembled using Flye assembler 2.7 (27), and polished with Illumina-data using Racon 1.4.20 (28). In silico MLST was performed using mlst 2.19.0 (28) and SCCmec typing using SCCmecFinder 1.2 (29). Acquired resistance genes were predicted using ResFinder 4.7.2 (30). Plasmid classification and typing were performed using COPLA (31) and PLSDB (32). Virulence genes were detected using abricate (33) against the VirulenceFactor DataBase (VFDB) (34) with identity ≥60% and coverage ≥50%.