Background on Salmonella Paratyphi B, variant Java
Salmonella Paratyphi B and Java are biovars of common serotype 1,4,[5], 12:b:1,2 which respectively cause human paratyphoid fever and gastroenteritis [1]. The d-tartrate fermenting variant S. enterica subspecies enterica serovar Paratyphi B dT+ (formerly called and referred to in this report as S. Java) is thought to be less virulent to humans than the non d-tartrate fermenting variant, although cases of invasive infection have been reported [1,5].

S. Java is a complex organism with a range of genetically distinct clonal lineages. Unlike ‘classical’ S. Paratyphi B, S. Java has an animal reservoir and has caused substantive outbreaks as a result of the contamination of food products. For example, in France in 2003 when 273 people were infected and one death was reported in an outbreak associated with contaminated goat’s milk cheese [5]. More recently, two distinct clonal lines characterised by resistance to antimicrobial drugs have been widely distributed. One of these was frequently found in poultry production, in the Netherlands, between 2000 and 2004 [4-7], and caused infections in humans in several countries in Europe as a result of the importation of infected poultry from the Netherlands [8]. The second clonal line, characterised by possession of Salmonella Genomic Island 1 (SGI1), has become widely distributed world-wide. These strains are generally resistant to ampicillin, chloramphenicol, streptomycin, spectinomycin, sulphonamides and tetracyclines [3,9,10] and have been associated with aquaria, in particular tropical fish aquariums [11], although outbreaks in cattle have also been reported [12].

Outbreak description
On 15 August 2007, Enter-net, the international surveillance network for the enteric infections Salmonella and VTEC O157, circulated a PFGE profile of S. Java outbreak strain in response to a cluster of cases first identified in Sweden on 10 August 2007. On 5 December 2007, the United Kingdom (UK) issued another urgent inquiry after an investigation of 20 cases of S. Java phage type 3b var 9; the majority of isolates were sensitive to all antimicrobial drugs at the levels used in the HPA Laboratory of Enteric Pathogens.

The rapid inquiry sent through the network requested Member States (MS) to check the occurrence of cases with a strain matching the PFGE pattern seen in the UK. As of 18 December, the results of this inquiry are the following:

• In Sweden, a total of 172 cases have been identified in a large outbreak between July and September 2007. The strain identified in this year demonstrated a PFGE pattern identical to the strains isolated from cases in the previous year (September-October, 2006). This PFGE profile has been designated SPTJXB.0001 in accordance with the Pulse-Net Europe [13] designations for PFGE profile types. An epidemiological investigation showed a strong association with imported baby spinach. Spinach samples were taken by the Swedish food authorities but Salmonella was not detected in any of the food samples. In the Swedish outbreak, more than 40% of the first 116 cases were hospitalized and no deaths have been reported. As a result of the Swedish investigation linking the outbreak with imported baby spinach, an alert was issued within the Rapid Alert System for Food and Feed (RASFF) on 24 August 2007 by the Swedish authorities. Five additional cases have become ill in November 2007.

• In Denmark, a total of 15 cases with SPTJXB.0001 were identified between August and November 2007; the PFGE pattern is pending for another four recent cases. In 2006, seven cases with SPTJXB.0001 were identified in September-October. Interviews of a subset of the 2007 cases have so far not confirmed baby spinach as the vehicle of infection nor led to alternative hypothesis.

• The UK is currently investigating 22 cases identified since 1 November 2007 (including three cases from Scotland and two believed to be from secondary spread). The PFGE pattern has shown to be indistinguishable from that of SPTJXB.0001. Twelve of the 17 primary cases in England have been interviewed to date to develop a hypothesis for disease transmission. It is notable that 11 of the 12 cases reported the consumption of salad vegetables purchased from a number of retailing and catering outlets. No single leaf type has been clearly identified. The frequencies of consumption of other foods were all markedly lower, and the investigation has ruled out fruit, fish, shellfish, dairy products, all meats, contact with animals and travel abroad. No new cases have been identified in December, but earlier cases occurring in August, September and October have been reported yet their exact numbers have not been established yet.

• In Finland, one case has been identified in August. Importantly, this case became ill after visiting Sweden. The strain isolated had the SPTJXB.0001 profile. No information is available on the food exposures of this case.

• In Norway, 10 cases with SPTJXB.0001 have been identified in August. Eight of these had also visited Sweden before becoming ill. No information on the food exposures of the ten cases is available.

• In the Netherlands, two cases with SPTJXB.0001 have been identified in October 2007. One of them had travel history to Bosnia-Herzegovina.

• In the United States, one case with an identical PFGE pattern to SPTJXB.0001 has been reported in August of 2007. This case will be interviewed to learn of any travel to Europe.

ECDC risk assessment
In an effort to respond to the ongoing outbreak and the increasing numbers of cases of acute gastroenteritis infected with S. Java strain with SPTJXB.0001 profile, the ECDC produced a risk assessment for the European Commission with the following conclusions:

Cases have continued to occur since the summer of 2007 and microbiological evidence of the SPTJXB.0001 profile points towards a dispersed multinational outbreak with a continuous and sustained risk to human health in Europe. Of the over 300 cases of S. Java that have been reported in 2007, many are likely to be linked to this ongoing outbreak.

Based upon epidemiological evidence from Sweden, imported baby spinach appears to be the most likely vehicle, yet no microbiological confirmation of this has been obtained.

As there is a potential to have more cases linked to this outbreak in the future, efforts to confirm a food source through microbiological investigations should be encouraged via the collaboration between national public health and food safety authorities, especially in Sweden, the UK, and Denmark.

Summary and conclusion
In 2007, 354 S. Java cases have been reported to ECDC by 11 MS (Figure 1). 228 of these cases (as of December 18, 2007) have an indistinguishable PFGE pattern, designated SPTJXB.0001, and are therefore possibly linked to the multinational outbreak (Figure 2). Unfortunately microbiological investigations have failed to confirm the incriminated vehicle of the Swedish cases.

Evidence for a possible vehicle in this multi-national outbreak is based on a case control study conducted in Sweden in response to their late summer surge in cases. There, the cases were strongly associated with the consumption of imported baby spinach. The spinach was initially sold as a separate product, yet later distributed as a part of a mixed salad product and served raw (whereas normal sized spinach is typically not used in salads but merely in warm meals that undergo heat treatment). Salmonella, however, was not detected in baby spinach samples taken by the Swedish food authorities. A RASFF was issued by the Swedish authorities in August of 2007 and subsequently a substantial decrease in cases was observed in Sweden. In spite of this transient decrease, cases continue to be reported yet at a lower frequency.

The UK epidemiologic investigation results are consistent with salad vegetables being a possible vehicle of infection, yet no single product or outlet was implicated and cases could not identify which type of leaf vegetable they had eaten. More information is clearly needed on the results of random sampling of imported spinach products in order to learn if the risk continues from spinach or is now in a mixed salad vehicle. An in-depth traceback analysis and analysis of supply channels is essential and would greatly enhance the efficiency of any microbiological testing of imported salads by focussing on the most ‘at risk’ products. ECDC will further explore this question.

In addition to further exploring the spinach and mixed salad as the possible vehicle, enhanced efforts to identify a common source for this S. Java SPTJXB.0001 outbreak are needed. The authors hope to raise awareness through this article and by sending urgent inquiries and their updates via ECDC and thus contribute to the increase in frequency of sampling suspected food sources and collaboration between national public health and food authorities. This may, in consequence, add to the likelihood of identifying a common source of the outbreak. A more rapid response to the investigation into S. Java cases should minimize recall bias and further improve the likelihood of detecting the organism in suspected products and thus help to solve such kinds of investigations.