We read with particular interest the article published recently by Merino et al., entitled “Serratia marcescens bacteraemia outbreak in haemodialysis patients with tunnelled catheters due to colonisation of antiseptic solution. Experience in 4 hospitals”,1 to which we would like to contribute our experience.

As reported in the aforementioned article,1 between December 2014 and January 2015, we recorded several cases of catheter-related bacteraemia (CRB) due to Serratia marcescens (S. marcescens) in our haemodialysis (HD) unit. On 19 December 2014, the Spanish Agency of Medicines and Medical Devices (AEMPS) ordered the withdrawal from the market of BohmClorh® chlorhexidine 2% aqueous antiseptic solution for healthy skin as a result of the contamination of several lots with S. marcescens.2 Several subsequent warnings were issued extending the restriction of its use to the remaining lots and presentations and, finally, on 9 January 2015, the master formulae developed by the company were withdrawn.3

Between 1 December 2014 and 16 January 2015, we recorded 14 cases of blood culture-confirmed CRB due to S. marcescens, 26.9% of the patients in our unit; all had been or were catheter users.

Suspicion of an epidemic outbreak arose after the appearance in the same week of 4 cases of bacteraemia caused by a microorganism that was unusual in the HD unit, in patients on different shifts, with differing dialysis stations and healthcare staff.

The Medical Prevention Service was informed and the corresponding protocol was carried out, reviewing the asepsis measures for catheter handling, and investigating the unit consumables and non-consumables and healthcare staff as possible foci and vectors of transmission.

A total of 9 men and 5 women with mean age 64±20 years were affected. Twelve patients (85.7%) were catheter users: 11 had a tunnelled catheter while only one had a temporary catheter; the remaining two recently had temporary catheters. A total of 26 episodes were recorded (mean of 1.85 episodes per patient), considering these as the onset of general malaise, hypotension and fever during the dialysis session, with blood cultures positive for S. marcescens. Two patients had up to 4 episodes of bacteraemia, one patient had 3 episodes, 4 patients had 2 episodes, and 7 patients had a single episode before the causal agent was identified. The catheter had to be removed in 5 patients, all of who were found to have infection with the same epidemic agent. Half the patients required hospitalisation for sepsis, with an average stay of 3.3 days. There were no deaths. One patient died due to CRB associated with a different microorganism.

In the early cases, patients were empirically administered intravenous vancomycin (1g post-dialysis) and gentamicin (at a dose of 1mg/kg/weight post-dialysis) until the blood culture results became available. Ciprofloxacin treatment was then instigated in all patients according to antibiotic susceptibility testing (AST), and the catheter was locked with the same antibiotic solution; one case required carbapenems due to poor clinical progress, and no antibiotic resistance was noted in the AST.

Pharyngeal and perianal exudates were systematically collected from all patients in the unit. Only one of the patients was a pharyngeal carrier. In two of the controls, the pharyngeal exudate was positive for S. marcescens, but for a strain other than the original one in the outbreak.

There were cases in other hospital units and other Spanish hospitals.4,5 Once the outbreak had been notified by AEMPS, the antiseptic product was cultured and confirmed as the reservoir of the epidemic. No further cases were recorded after the contaminated lots had been withdrawn.

Most CRB in HD patients is caused by Gram-positive bacteria.6S. marcescens is a gram-negative bacillus found predominantly in humid conditions; it is pathogenic for man and causes outbreaks of resistant infections, especially in immunosuppressed patients.6

The general rule in CRB is to remove the catheter, but conservative treatment is accepted in HD patients despite evidence of serious infection, as a measure to preserve the vascular access.1,7

In our experience, AST-directed treatment and catheter locking provided a good outcome, although the delay in identifying the source and probable formation of biofilms meant that a significant number of catheters had to be removed due to the high percentage of recurrences.

Although the finding of S. marcescens in HD units is rare, and identification of the reservoir in an antiseptic product is paradoxical, outbreaks have previously been described.8,9 A methodological search for the causal agent and appropriate preventive strategies should therefore be instigated at an early stage when an epidemic outbreak is suspected. An initiative to record the experience in the HD units of all the affected hospitals would be interesting.