Treatment of healthcare workers carrying SA t408, and re-implementing and re-emphasising hygienic measures were effective to control the outbreak of SA t408 among neonates in the Jeroen Bosch Hospital. Rapid typing tools are indispensable to support adequate infection control measures. No new cases have been reported since November 2011. However, increased awareness is maintained in our setting to identify new cases. Furthermore, it is recommended that all fusidic acid resistant strains isolated from neonates are typed and analysed for possible clustering with other isolates in our hospital.
Routine surveillance of MSSA is not common practice in The Netherlands, so the epidemiology of MSSA and of specific methicillin susceptible isolates, such as t408, is largely unknown. This MSSA type, t408, is possibly linked to the EEFIC clone, which has caused outbreaks at neonatal wards causing BI and SSSS in several European hospitals [8, 10, 12]. EEFIC is most commonly seen as t171 but has also has been described as single locus variants t408, t659, t874, and t875 . All belong to MLVA complex 123 [pers. comm., de Neeling].
All newly identified SA t408 strains were fusidic acid resistant (EUCAST criteria) with MIC values 3–8 mg/L, while the non-t408 SA strains were fusidic acid sensitive. Considering the number of fusidin resistant (R) S aureus isolated from 2007 until 2011 in neonates in our hospital (total of 188 MSSA S aureus isolates on unique patients of max. 31 days old), we noticed a significant increase starting from the third quarter in 2010. In addition, the here described SA t408 strains were MecA-, PVL-, ETA-, ETB+, and ETAD-, like most other EEFIC strains[7–10, 13, 14].
Recently, Rijnders et al.  studied this EEFIC clone in the South of the Netherlands by examining skin and soft tissue infections (SSTI) at GPs in 2007–2008. Of all participating SSTI patients seen by GPs in the southern part of the Netherlands, approximately 2.5% carried t408. This finding leaves room to speculate whether the cases described here were either related to a local outbreak, or a representative sample of the MSSA circulating in the community at the time. On the other hand, re-implementing stricter hygienic protocols did have an impact on the amount of new cases, suggesting a local outbreak was controlled.
Moreover, no overlap in hospital stay between was seen for some of the clustered cases. Patient P (Table 1) was found in November of 2011, three months after the last case of t408 was found. Patient A - the first case to be spa typed - had an unrelated Raman-type compared to the cases that were typed later.
Unfortunately, we have no data on frequencies of circulating methicillin susceptible SA spa types in the area of our hospital. This study shows spa type t408 in nine out of sixteen neonatal cases, in contrast to surveillance cultures among HCW which yielded only three t408 among 131 isolated MSSA. It is possible that t408 is the predominant strain amongst neonates in our area.
Raman typing appears to be a convenient method for epidemiological typing in an outbreak situation. It has superior discriminative power compared to spa-typing, which was not specific enough to distinguish between patient A (Table 1) and the cluster that was found. Patient A was the first typed t408 patient and had an unrelated Raman type. Raman-typing is highly specific, is fast (within an hour) with high throughput, and is an easy-to-use method for outbreak situations . In this study both spa-typing and Raman typing were used. Ideally, real time results identifying the incriminated strain would be most desirable to adequately install infection control measures. Spa-typing is more laborious than Raman typing and has possibly less discriminative power for this typical application (i.e. S. aureus t408). Raman typing is an expensive technique and specific equipment is needed. An important drawback of Raman typing is that the output data are not easily comparable between different databases since the output data can not be translated into exact numbers (as in spa typing). Therefore, Raman typing is not suitable for large scale outbreak settings or (inter)national surveillance.
From the observation at the operation theatre we concluded that it was highly unlikely that a member of the operation team could contaminate the neonate, since there is no direct skin contact between the operating team and the neonate. However possible sources of contamination existed because the resuscitation table was not properly cleaned, materials from previous caesareans were still present at the table, and a cleaning protocol for the resuscitation table did not exist.
Our advice was to devise a protocol for the cleaning of the table, and to write a protocol for twin deliveries, since this was missing.
During the observation of the neonatal ward we concluded that the resuscitation table, emergency trolley, and the incubators were not cleaned properly. Therefore we took environmental swabs of these surfaces as well. We also concluded that in the storage room no clear division between sterile and non-sterile products was maintained. We also noticed that a clear cleaning protocol was missing. These points were taken to the relevant wards/nurses and the hygienic protocols were sharpened/re-implemented.
Because most of our IB cases were caused by MSSA strains other than t408, we are interested to do a follow-up study where we will not only focus on bullous impetigo, but also on other impetigo cases seen in child day-care centres and GPs, to get a better understanding of the epidemiology of virulent MSSA strains amongst infants and older children.
Our outbreak peaked early summer in May, and lasted from March until August, with one last case in November. It seems unlikely that this outbreak can be linked to the yearly increase in impetigo cases that is seen in day-care centres and nurseries in late summer, since our outbreak didn’t peak in late summer [1, 2, 17].