Changing Pattern of Human Listeriosis, England and Wales, 2001–2004

Disease has reemerged, mainly in patients ≥60 years of age with bacteremia.


Materials and Methods
The Health Protection Agency Centre for Infections (CFI) coordinates routine surveillance of human listeriosis in England and Wales. Case ascertainment is by the voluntary reporting of laboratory-diagnosed cases from microbiology laboratories through an electronic reporting system, or by referral of cultures for identification and subtyping. Epidemiologic and microbiologic data from both systems are combined, checked for duplication, and stored in a database. Additional demographic and clinical data are sought from the responsible consultant medical microbiologists and local health protection teams with a standard questionnaire.
For surveillance purposes, a patient with listeriosis was defined as one with a compatible illness from whom L. monocytogenes was isolated from a normally sterile site, Changing Pattern of Human Listeriosis, England and Wales, [2001][2002][2003][2004] usually blood or cerebrospinal fluid (CSF). Cases were categorized as pregnancy associated (all maternal-fetal patients and neonatal patients; a mother-baby pair was considered 1 case-patient) and nonpregnancy associated (in a patient >1 month of age). Case-patients were categorized further into those with CNS infections (L. monocytogenes isolated from CSF or brain tissue, clinical evidence of infection of this organ, or both); bacteremia in the absence of CNS infections (L. monocytogenes isolated from blood but not from CNS and without clinical evidence of CNS infection); and other conditions not included in the previous 2 categories. Data manipulation was undertaken in Microsoft Access 2003 (Microsoft Corporation, Redmond, WA, USA) and MapInfo version 8.0 (MapInfo Corporation, Troy, NY, USA). Ethnicity (categorized as ethnic or nonethnic) was assigned by using patients' names (surname and first name when available); patients' ages were grouped into 10-year bands. Indexes of deprivation for England (2004 [13]) and Wales (2005 [14]), ranked and arranged into quintiles (1 = most deprived and 5 = least deprived areas) and linked to patients postal codes, were used as an approximate marker for patients' socioeconomic status. Patients' postal codes were also used as a marker for patients' residency. Internet searches were used to determine if residential care homes were situated in that postal code area or whether the housing was purely residential.
Data analysis was performed with Microsoft Excel, EpiInfo version 6.04b (Centers for Disease Control and Prevention, Atlanta, GA, USA) and Stata version 8.2 (StataCorp, College Station, TX, USA). Age-specific denominator data from 1990 to 2004 were obtained from the Office for National Statistics. Relative proportions and changes in relative proportions with time were compared by using the χ 2 test and the χ 2 test for trend, respectively. Point estimates of relative risks (RRs) with accompanying 95% confidence intervals (CIs) and significance tests were also calculated.

All Reported Cases
From January 1, 1990, to December 31, 2004, a total of 1,933 reported cases of human listeriosis in England and Wales fulfilled the case definition. Of these 1,377 (71%) were reported through the electronic surveillance system, 1,592 (82%) by isolate referral and 1,068 (55%) by both means. During the study period, the proportion of isolates referred did not change (χ 2 for trend p = 0.94); the proportion of electronic reports received increased slightly (χ 2 for trend p = 0.04).

Clustered Cases
Epidemiologic and subtyping analysis identified 10 clusters of cases, which affected 60 patients and likely reflected common-source outbreaks (15,16); these are summarized in Table 1. When these cases were excluded, a significant increase in disease in 2001-2004 compared with 1990-2000 remained (RR 1.34, 95% CI 1.26-1.42, p<0.001). Subsequent analysis is confined to 1,873 sporadic cases unless otherwise indicated.

Trends in Sporadic Cases
In 1990, sporadic nonpregnancy-associated listeriosis accounted for 80% of the 114 cases reported, and in 2004 for 90% of the 205 cases reported (χ 2 for trend p<0.001; Figure 1). A total of 510 (44%) of 1,155 of the nonpregnancy-associated patients and 29 (10%) of 287 of the sporadic pregnancy-associated patients died.
Data on sex were available for 1,542 (99%) of the 1,543 nonpregnancy-associated patients for whom age was also available.  Patients' postal codes were available for 634 (58%) of 1,102 nonpregnancy-associated patients >60 years, and indexes of deprivation were determined for 563 (89%). The proportion of patients who fell into the quintiles of deprivation 1-5 did not change during the surveillance period (χ 2 for trend p = 0.57, 0.69, 0.64, 0.05, and 0.14, respectively).
Internet searches of the areas covered by the postal codes of the 634 nonpregnancy-associated patients >60 years of age showed that, when genuine postal codes were supplied (628, 99%), most (580, 92%) did not contain a residential care home. This proportion did not differ

Discussion
Routine surveillance of human listeriosis in England and Wales showed an upsurge in cases such that the annual incidence is now comparable with other European countries with higher incidence (17). The clinical manifestations have also changed: bacteremia in older patients without CNS involvement predominates. Several confounding factors could explain the increase in cases and changes in signs and symptoms.
Changes in reporting or referral could have accounted for the observed increase in incidence. The surveillance of listeriosis in England and Wales is passive, and such systems are prone to both underascertainment and pseudooutbreaks following increased interest in the public health community. Although reporting artifacts cannot be excluded, we are unaware of increased interest in listeriosis from 2001 onwards. Furthermore, reporting and referrals did not change enough to explain the increase. Improvements in laboratory methods (especially in the isolation of L. monocytogenes from blood) or changes in local clinical practice (e.g., more detailed investigations of patients with acute febrile illness seeking primary care) might explain the increase in cases diagnosed or the altered clinical manifestations. We are unaware of substantial changes in blood culture techniques used in England and Wales in the past decade that would increase the diagnosis of listeriosis. Furthermore, although the introduction of mandatory reporting of methicillin-resistant Staphylococcus aureus bacteremia in England in 2001 has led to an increase in blood cultures being taken, this is insufficient to explain the increase or shift in clinical manifestations described here (18,19). Further evidence that the increase was not due to improved diagnostics is the absence of statistically significant increase in the isolation of L. monocytogenes from blood cultures from patients with CNS infections or from pregnancy-associated patients.
Demographic changes in the population might have resulted in an overrepresentation of patients from particular age groups without a true increase in risk. Life expectancy in the United Kingdom is increasing; therefore, an increase in listeriosis in older patients is likely to occur. However, calculations controlling for the changing age structure in England and Wales during the surveillance period generates a consistent increase in risk among those >60 years of age. Medical advances have resulted in the UK population's surviving for longer with chronic conditions (20) with a likely increased susceptibility to listeriosis. While the denominator data required to examine such changes in detail are unavailable, changes would be unlikely to result in an almost 3-fold increase in a single patient age group in a short period without a concomitant increase in younger patients with similar underlying conditions. Changes in the pathogenicity of L. monocytogenes might explain the change in disease manifestations. However, the increase has been due to multiple subtypes, which makes this unlikely. Furthermore, since the upsurge was confined to a restricted patient age group, it is more likely to reflect increased incidence through higher exposure that accompanies behavioral changes. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 12, No. 9, September 2006  Having examined the most plausible sources of bias, we believe that the observed upsurge and altered clinical manifestations are genuine. Indeed, historical data suggest that the current picture merely represents a continued shift in the epidemiology and clinical manifestations of L. monocytogenes infection in England and Wales (Table 3 (4,21).
The routine epidemiologic and microbiologic data collected for cases of listeriosis in England and Wales are not exhaustive; therefore, our retrospective examination of the factors that have contributed to this upsurge is preliminary. Nevertheless, we have demonstrated that the upsurge is independent of sex; regional, seasonal, ethnic, or socioeconomic differences; underlying conditions; or L. monocytogenes subtypes. Furthermore, most older patients in the surveillance period did not reside in care homes and were therefore unlikely to have changed exposure to institutional catering in such settings. UK food consumption/expenditure data also suggest that no major shift in the consumption of major food groups by the older population has occurred in recent years to explain the increase (22).
Investigations are continuing to establish the causes of the increase and include application of discriminatory subtyping of L. monocytogenes isolates, coupled with the collection of standardized clinical and epidemiologic data for all patients. Hopefully, such steps will facilitate outbreak detection and help identify their cause, as well as enable investigations of factors specific to L. monocytogenes subtypes among sporadic cases. However, analytical epidemi-ology (including case-control studies) and molecular fingerprinting of isolates have not always successfully identified the appropriate interventions to control outbreakassociated and sporadic listeriosis, which suggests that new approaches to investigation are required. Therefore, in the absence of risk factors for listeriosis in this emerging at-risk sector of the population, dietary advice on the avoidance of high-risk foods should be provided routinely to the elderly and immunocompromised and not just pregnant women.