Nosocomial Spread of a Staphylococcus hominis subsp. novobiosepticus Strain Causing Sepsis in a Neonatal Intensive Care Unit

Abstract

From 2002 to 2003, 32 isolates of Staphylococcus hominis subsp. novobiosepticus (SHN) were recovered from 21 patients, 18 of whom were neonates, with 13 considered to have late-onset SHN sepsis. All isolates from neonates had an indistinguishable pulsed-field gel electrophoresis pattern. Our data support SHN as an important nosocomial pathogen in neonates.

Coagulase-negative staphylococci (CoNS) are now recognized as a major cause of nosocomial infections in neonatal intensive care units (NICUs) (8) and are responsible for 48% of late-onset sepsis among very-low-birth-weight neonates (20). Although Staphylococcus epidermidis causes 60 to 93% of CoNS bloodstream infections, several other CoNS species are reported to cause disease in infants (10, 11, 18). Recently, a novel subspecies of Staphylococcus hominis, S. hominis subsp. novobiosepticus (SHN), was isolated from blood cultures and other clinical specimens (6, 12). The name derives from the combination of novobio, pertaining to the property of novobiocin resistance, and septicus, pertaining to the ability to cause sepsis. After isolating SHN from several blood cultures at the NICU during 2002 and 2003, we undertook a retrospective study in our 1,300-bed tertiary care institution to determine the prevalence of this subspecies, the clinical significance of the isolates, and whether the infections were clonal in origin.

Clinical records of patients from whom SHN had been isolated were reviewed. Diagnosis of sepsis was based on the assessment of the attending physician. For NICU patients, true infection was defined either by isolation of SHN from two separate blood cultures or by a single SHN-positive blood culture and an elevated serum level of C-reactive protein (≥2 mg/dl). Blood samples were inoculated and processed in an automated system (BacT/Alert [BioMérieux, Durham, NC] or BACTEC 9240 [BD Diagnostics, Sparks, MD]). The type strain of SHN (ATCC 700236) was used as a reference standard throughout the investigations. Identification and susceptibility testing were performed using conventional methods and the Wider system (Soria Melguizo, Madrid, Spain) (3, 16). A zone of inhibition measuring ≤15 mm in Mueller-Hinton agar or ≤11 mm on Trypticase soy agar plates was considered indicative of novobiocin resistance (12). All isolates of SHN underwent confirmatory PCR analysis for the mecA gene (7). To confirm the identification of SHN, broad-spectrum primers were used for sequencing the 5′ ends of both strands of the 16S rRNA gene (2, 6) using an ABI Prism 3100 genetic analyzer (Applied Biosystems, Foster City, CA). Further molecular characterization of SHN isolates was performed by pulsed-field gel electrophoresis (PFGE) following digestion of DNA extracts with SmaI (5).

During 2002 and 2003, 32 isolates of SHN were recovered from 21 patients. Twenty-three isolates were from blood cultures, six were from catheters, one was from cerebrospinal fluid (CSF), one was from a wound, and one was from external ear fluid. All 21 patients yielded an SHN-positive blood culture. All isolates were distinguished from S. hominis subsp. hominis on the basis of failure to produce acid aerobically from d-trehalose and resistance to novobiocin. Partial DNA sequencing of the 16S rRNA gene from four clinical isolates, representing three different PFGE patterns (see below), and the ATCC type strain demonstrated 100% homology (6). Overall, of 21 isolates of SHN from blood (1 per patient), 21 (100%) were resistant to penicillin, oxacillin, and erythromycin while 20 (95.2%) were resistant to clindamycin and gentamicin, 19 (90.5%) were resistant to tetracycline, 6 (28.6%) were resistant to trimethoprim-sulfamethoxazole, 5 (23.8%) were resistant to chloramphenicol, and 2 (9.5%) were resistant to ciprofloxacin. All 21 isolates were resistant to nalidixic acid and susceptible to vancomycin. PCR analysis confirmed that all strains possessed a mecA gene homologue.

Of 21 patients with SHN bacteremia, 18 were neonates, 1 was a 13-year-old boy, and 2 were adults admitted to the obstetric and internal medicine wards, respectively. All patients had single episodes of SHN bacteremia. The demographic and clinical characteristics of the neonates are summarized in Table 1. According to the definition of sepsis in neonates, 13 of these episodes were considered definite cases of SHN infection. All of these patients were hospitalized from birth and were considered to have late-onset sepsis (occurring at >72 h of age), with a median age at presentation of 9 days (range, 5 to 65 days). All the infants had received previous antimicrobial therapy with ampicillin and gentamicin and had intravascular catheters in place at the time of their bacteremia. All but cases 4 and 20 received treatment with vancomycin and amikacin. At the same time as the SHN infection, case 20 had another clinical episode of bacteremia with an extended-spectrum β-lactamase-producing strain of Klebsiella pneumoniae and received several regimens of antimicrobial treatment. None of the patients died of causes related to their SHN sepsis.

TABLE 1.

Demographic and clinical characteristics of neonates with SHN bacteremia

Case no.	Sex	Gestational age (wk)	Birth wt (g)	Date of bacteremia (mo/day/yr)	Age at bacteremia (days)	Other source(s) of SHN isolation	Underlying condition(s)	Clinical significancea
1	Male	29	1,251	2/28/2002	9	None	RDSb	No
2	Female	31	1,395	3/8/2002	6	Catheter	RDS	Yes
3	Male	27	1,020	4/8/2002	11	None	RDS	No
4	Female	27	900	7/2/2002	9	None	RDS	Yes
5	Male	32	1,650	8/5/2002	10	None	None	Yes
6	Male	35	2,340	8/8/2002	35	Catheter	Necrotizing enterocolitis	Yes
7	Female	41	1,750	8/28/2002	7	None	None	Yes
8	Male	35	1,209	9/6/2002	11	CSF	Malformative syndrome	Yes
9	Female	36	3,000	9/10/2002	9	None	Omphalocele	No
10	Female	33	1,830	9/30/2002	7	Catheter, wound	RDS, pulmonary hypoplasia	Yes
11	Male	37	2,120	10/30/2002	65	None	Gastroschisis	Yes
12	Female	36	2,950	10/30/2002	29	None	Congenital heart disease	No
16	Male	26	800	9/22/2003	120	None	Bronchopulmonary dysplasia	No
17	Male	26	950	9/29/2003	9	Catheter, ear fluid	RDS	Yes
18	Female	25	711	10/17/2003	5	Catheter	RDS	Yes
19	Female	38	2,300	10/17/2003	16	None	Intestinal obstruction	Yes
20	Male	28	520	10/20/2003	12	Catheter	RDS	Yes
21	Female	29	1,258	10/20/2003	8	None	RDS	Yes

^aClinical significance was assessed as described in the text.

^bRDS, respiratory distress syndrome.

Twenty-three isolates of SHN (21 from blood, 1 from CSF, and 1 from a catheter), corresponding to 21 patients, and one type culture (ATCC 700236) were available for PFGE typing (data not shown). Twenty-one isolates from the 18 neonates and 1 13-year-old boy (case 15) had an indistinguishable macrorestriction pattern (pattern A) that differed by three bands from the patterns of the two adult patients (cases 13 and 14). All isolates with PFGE pattern A were susceptible to ciprofloxacin and resistant to tetracycline. Most isolates of SHN from the neonatology department were obtained during two separate periods of time: eight cases between August and September, 2002, and six cases between September and October, 2003 (Table 1). The second period was coincidental with an outbreak of infections in this department due to extended-spectrum β-lactamase-producing K. pneumoniae.

After the initial description of SHN in 1998 (12), to our knowledge, this is the first clinical report implicating this microorganism as the cause of bacteremia in hospitalized patients and particularly among patients from a NICU. As in other studies on neonatal CoNS infections that have demonstrated significant morbidity but a low rate of mortality, there were no deaths associated with SHN sepsis at our institution (15, 23). The incidence and types of infection associated with SHN in hospitalized patients are not known as, like others, our laboratory does not routinely identify CoNS to the species level. Although there is no information on the prevalence of colonization with SHN, Kloos et al. (12) suggest that if SHN is a resident on human skin, it probably persists in very small numbers and requires enrichment for detection. In our study, 6/13 (46%) neonates with clinically significant bacteremia had proven catheter-related bloodstream infections. CoNS account for a significant proportion of nosocomial bacteremia cases related to the insertion and maintenance of intravascular catheters (1, 22). It is likely that more cases of neonatal bacteremia in our study were also catheter related, but catheters were not processed for culture in all instances.

An important characteristic of SHN is its pattern of multidrug resistance, including resistance to oxacillin. This has important clinical implications because the alternative to oxacillin for treatment of these patients is administration of vancomycin, the overuse of which has implications for the continued emergence of vancomycin resistance in both CoNS and Staphylococcus aureus (4, 19, 21).

Molecular epidemiology showed that a single clone of SHN caused sepsis in at least 13 neonates in the NICU during the 2-year study period. Surprisingly, following the first outbreak in August 2002, after an interlude of about a year, the same strain was associated with another outbreak in September 2003. Empirical antibiotic therapy with the combination of ampicillin and gentamicin is likely to have favored selection of this multidrug-resistant SHN clone, which subsequently became endemic within our institution (14). Although we did not investigate the reservoir and mode of transmission, it has been postulated that the infants themselves are the reservoirs of endemic strains of staphylococci, with organism transmission occurring through contact with the hands of health care workers (4). Furthermore, other reports have shown that CoNS isolates from the nasopharynges and hands of health care workers are often genetically related to the organisms that colonize and/or cause disease in neonates, confirming the importance of health care workers in the nosocomial transmission of CoNS in the NICU setting (9, 17).

In summary, our data support that SHN is a significant cause of nosocomial bacteremia in the NICU and that, like other species of CoNS (9, 13), isogenic clones can persist during long periods of time and be transmitted between neonates.

Nucleotide sequence accession number.

The partial 16S rRNA sequence of the clinical isolate of SHN from blood corresponding to case number 2 was deposited in GenBank under accession number DQ056840.