Combining Whole Genome Sequencing Data from Human and Non-Human Sources: Tackling Listeria monocytogenes Outbreaks.

Listeria monocytogenes (Lm) is ubiquitous in nature and known for its ability to contaminate foods during production processes. Near real-time monitoring of whole genome sequences from food and human isolates, complemented with epidemiological data, has been used in the Netherlands since 2019 to increase the speed and success rate of source finding in the case of (active) clusters. Nine clusters with 4 to 19 human cases investigated between January 2019 and May 2023 are described. Fish production sites were most often linked to outbreaks of listeriosis (six clusters), though other types of food businesses can face similar Lm problems, as the production processes and procedures determine risk. The results showed that low levels of Lm in food samples can still be linked to disease. Therefore, the investigation of a cluster of cases and deployment of the precautionary principle helps to focus on safe food and to prevent further cases. Good practice of environmental monitoring within a food business allows early detection of potential issues with food safety and helps food businesses to take appropriate measures such as cleaning to prevent regrowth of Lm and thus future outbreaks.

Seroprevalence of Toxoplasma gondii and associated risk factors for infection in the Netherlands: third cross-sectional national study.

A third nationally representative serosurvey was performed to study the changes in Toxoplasma gondii (T. gondii) seroprevalence in the Netherlands over a 20-year time span and to identify and confirm risk factors for acquired toxoplasmosis. This cross-sectional study (conducted in 2016/2017) was designed similarly to the previous two studies (1995/1996 and 2006/2007) and included a questionnaire and serum sampling among Dutch residents. Factors associated with seropositivity for T. gondii were determined using multivariable analysis of the questionnaire-derived data. The earlier observed decrease in T. gondii seroprevalence between 1995/1996 and 2006/2007 (from 40.5% to 26.0%) did not continue into 2016/2017 (29.9%). Similarly to the previous studies, the seroprevalence increased with age and varied among regions. In all studies, higher T. gondii seropositivity was associated with increasing age, lower educational level, not living in the Southeast, and eating raw or semi-cooked pork. The incidence of congenital toxoplasmosis was estimated at 1.3/1000 (95% CI 0.9-1.8) live-born children in 2017. As the seroprevalence of T. gondii in the Netherlands did not decrease over the last decade, an increase in public health awareness is needed and prevention measures may need to be taken to achieve a further reduction in T. gondii infections in the Netherlands.

Epidemiology of Pathogens Listed as Potential Bioterrorism Agents, the Netherlands, 2009‒2019.

We provide incidences (cases/10 million persons) in the Netherlands during 2009-2019 for pathogens listed as potential bioterrorism agents. We included pathogens from the highest categories of the European Medicines Agency or the US Centers for Disease Control and Prevention. Notifiable diseases and recently published data were used to calculate the average annual incidence. Coxiella burnetii had the highest incidence because of a Q fever epidemic during 2007-2010. Incidence then decreased to 10.8 cases/. Pathogens with an incidence >1 were Brucella spp. (2.5 cases), Francisella tularensis (1.3 cases), and Burkholderia pseudomallei (1.1 cases). Pathogens with an incidence <1 were hemorrhagic fever viruses (0.3 cases), Clostridium botulinum (0.2 cases), and Bacillus anthracis (0.1 cases). Variola major and Yersinia pestis were absent. The generally low incidences make it unlikely that ill-meaning persons can isolate these pathogens from natural sources in the Netherlands. However, the pathogens are stored in laboratories, underscoring the need for biosecurity measures.

Risk factors for acute toxoplasmosis in the Netherlands.

Toxoplasmosis caused by the protozoan parasite Toxoplasma gondii occurs worldwide. Infections range from asymptomatic to life-threatening. T. gondii infection is acquired either via bradyzoites in meat or via oocysts in the environment, but the relative importance of these path ways and the different sources remains unclear. In this study, possible risk factors for toxoplasmosis in the Netherlands were investigated. A case-control study was conducted including persons with recent infection and individuals with a negative test result for IgM and IgG for T. gondii between July 2016 and April 2021. A total of 48 cases and 50 controls completed the questionnaire. Food history and environmental exposure were compared using logistic regression. Consumption of different meats was found to be associated with recent infection. In the multivariable model, adjusted for age, gender, and pregnancy, consumption of large game meat (adjusted odds ratio (aOR) 8.2, 95% confidence interval 1.6-41.9) and sometimes (aOR 4.1, 1.1-15.3) or never (aOR 15.9, 2.2-115.5) washing hands before food preparation remained. These results emphasize the value of the advice to be careful with the consumption of raw and undercooked meat. Good hand hygiene could also be promoted in the prevention of T. gondii infection.

New insights into the epidemiology of Listeria monocytogenes - A cross-sectoral retrospective genomic analysis in the Netherlands (2010-2020).

Introduction: Listeriosis, caused by infection with Listeria monocytogenes (Lm), is a relatively rare but severe disease with one of the highest mortality rates among bacterial foodborne illnesses. A better understanding on the degree of Lm clustering, the temporal distribution of the clusters, and their association with the various food sources is expected to lead to improved source tracing and risk-based sampling. Methods: We investigated the genomic epidemiology of Lm in the Netherlands between 2010 and 2020 by analyzing whole-genome-sequencing (WGS) data of isolates from listerioss patients and food sources from nationwide integrated surveillance and monitoring. WGS data of 756 patient and 770 food/environmental isolates was assessed using core-genome multi-locus sequence typing (cgMLST) with Hamming distance as measure for pairwise distances. Associations of genotype with the epidemiological variables such as patient's age and gender, and systematic use of specific drugs were tested by multinomial logistic regressions. Genetic differentiation of the Lm within and between food categories was calculated based on allele frequencies at the 1701 cgMLST loci in each food category. Results: We confirmed previous results that some clonal complexes (CCs) are overrepresented among clinical isolates but could not identify any epidemiological risk factors. The main findings of this study include the observation of a very weak attribution of Lm types to food categories and a much better attribution to the producer level. In addition, we identified a high degree of temporal persistence of food, patient and mixed clusters, with more than half of the clusters spanning over more than 1 year and up to 10 years. Discussion: Taken together this would indicate that identifying persistent contamination in food production settings, and producers that process a wide variety of raw food produce, could significantly contribute to lowering the Lm disease burden.

Corrections to "Prevalence of persistent symptoms after treatment for lyme borreliosis: a prospective observational cohort study" [The Lancet Regional Health - Europe 6 (2021) 100142].

[This corrects the article DOI: 10.1016/j.lanepe.2021.100142.].

Prevalence of persistent symptoms after treatment for lyme borreliosis: A prospective observational cohort study.

BACKGROUND: Concerns about long-lasting symptoms attributed to Lyme borreliosis (LB) are widespread in the Western world, while such symptoms are highly prevalent in the general population. METHODS: In the largest prospective study to date, adults with physician-confirmed LB were included at the start of antibiotic treatment. Primary outcomes, prevalence of persistent symptoms and symptom severity, were assessed using three-monthly standardised questionnaires during one year. Persistent symptoms were defined as impaired scores for fatigue (CIS, subscale fatigue), cognitive impairment (CFQ) or pain (SF-36, subscale bodily pain) ≥6 months, with onset <6 months. Outcomes were compared with a longitudinal general population and a tick-bite cohort without LB as a reference. FINDINGS: Of 1135 LB patients (94•8% erythema migrans, 5•2% disseminated LB), 1084 fulfilled primary analysis criteria, as well as 1942 population and 1887 tick-bite controls. Overall prevalence of persistent symptoms in LB patients was 27•2% (95%CI, 24•7%-29•7%); 6•0% and 3•9% higher than in population (21•2%, 95%CI, 19•3%-23•1%; p < 0•0001) and tick-bite (23•3%, 95%CI 21•3%-25•3%; p = 0•016) cohorts, respectively. At 12 months, fatigue, cognitive impairment, and pain were significantly more severe in erythema migrans patients than in reference cohorts, while in disseminated LB patients, only pain was more severe. INTERPRETATION: In treated LB patients, persistent symptoms were significantly more prevalent and symptoms were more severe than in individuals without LB, although the background prevalence was substantial. This suggests an association, either direct or indirect, between persistent symptoms and LB in a relatively small subset of patients. FUNDING: ZonMw; Dutch Ministry of Health, Welfare and Sport.

Risk factors associated with the incidence of self-reported COVID-19-like illness: data from a web-based syndromic surveillance system in the Netherlands.

During the first wave of the severe acute respiratory syndrome-coronavirus-2 epidemic in the Netherlands, notifications consisted mostly of patients with relatively severe disease. To enable real-time monitoring of the incidence of mild coronavirus disease 2019 (COVID-19) - for which medical consultation might not be required - the Infectieradar web-based syndromic surveillance system was launched in mid-March 2020. Our aim was to quantify associations between Infectieradar participant characteristics and the incidence of self-reported COVID-19-like illness. Recruitment for this cohort study was via a web announcement. After registering, participants completed weekly questionnaires, reporting the occurrence of a set of symptoms. The incidence rate of COVID-19-like illness was estimated and multivariable Poisson regression used to estimate the relative risks associated with sociodemographic variables, lifestyle factors and pre-existing medical conditions. Between 17 March and 24 May 2020, 25 663 active participants were identified, who reported 7060 episodes of COVID-19-like illness over 131 404 person-weeks of follow-up. The incidence rate declined over the analysis period, consistent with the decline in notified cases. Male sex, age 65+ years and higher education were associated with a significantly lower COVID-19-like illness incidence rate (adjusted rate ratios (RRs) of 0.80 (95% CI 0.76-0.84), 0.77 (0.70-0.85), 0.84 (0.80-0.88), respectively) and the baseline characteristics ever-smoker, asthma, allergies, diabetes, chronic lung disease, cardiovascular disease and children in the household were associated with a higher incidence (RRs of 1.11 (1.04-1.19) to 1.69 (1.50-1.90)). Web-based syndromic surveillance has proven useful for monitoring the temporal trends in, and risk factors associated with, the incidence of mild disease. Increased relative risks observed for several patient factors could reflect a combination of exposure risk, susceptibility to infection and propensity to report symptoms.

Optimization of Notification Criteria for Shiga Toxin-Producing Escherichia coli Surveillance, the Netherlands.

We describe the consequences of 2 major changes in notification criteria for Shiga toxin-producing Escherichia coli surveillance in the Netherlands. The change to reporting acute, more severe infections appears to be a good compromise between workload, redundancy, and public health relevance, provided isolates remain available for typing and sequencing.

Incidence, clinical implications and impact on public health of infections with Shigella spp. and entero-invasive Escherichia coli (EIEC): results of a multicenter cross-sectional study in the Netherlands during 2016-2017.

BACKGROUND: Shigella spp. and entero-invasive E. coli (EIEC) use the same invasive mechanism to cause diarrheal diseases. Public health regulations apply only to Shigella spp. infections, but are hampered by the lack of simple methods to distinguish them from EIEC. In the last decades, molecular methods for detecting Shigella spp. and EIEC were implemented in medical microbiological laboratories (MMLs). However, shigellosis cases identified with molecular techniques alone are not notifiable in most countries. Our study investigates the impact of EIEC versus Shigella spp. infections and molecular diagnosed shigellosis versus culture confirmed shigellosis for re-examination of the rationale for the current public health regulations. METHODS: In this multicenter cross-sectional study, fecal samples of patients suspected for gastro-enteritis, referred to 15 MMLs in the Netherlands, were screened by PCR for Shigella spp. or EIEC. Samples were cultured to discriminate between the two pathogens. We compared risk factors, symptoms, severity of disease, secondary infections and socio-economic consequences for (i) culture-confirmed Shigella spp. versus culture-confirmed EIEC cases (ii) culture positive versus PCR positive only shigellosis cases. RESULTS: In 2016-2017, 777 PCR positive fecal samples with patient data were included, 254 of these were culture-confirmed shigellosis cases and 32 were culture-confirmed EIEC cases. EIEC cases were more likely to report ingestion of contaminated food and were less likely to be men who have sex with men (MSM). Both pathogens were shown to cause serious disease although differences in specific symptoms were observed. Culture-negative but PCR positive cases were more likely report travel or ingestion of contaminated food and were less likely to be MSM than culture-positive cases. Culture-negative cases were more likely to suffer from multiple symptoms. No differences in degree of secondary infections were observed between Shigella spp. and EIEC, and culture-negative and culture-positive cases. CONCLUSIONS: No convincing evidence was found to support the current guidelines that employs different measures based on species or detection method. Therefore, culture and molecular detection methods for Shigella spp. and EIEC should be considered equivalent for case definition and public health regulations regarding shigellosis. Differences were found regarding risks factors, indicating that different prevention strategies may be required.

An international outbreak of Salmonella enterica serotype Enteritidis linked to eggs from Poland: a microbiological and epidemiological study.

BACKGROUND: Salmonella spp are a major cause of food-borne outbreaks in Europe. We investigated a large multi-country outbreak of Salmonella enterica serotype Enteritidis in the EU and European Economic Area (EEA). METHODS: A confirmed case was defined as a laboratory-confirmed infection with the outbreak strains of S Enteritidis based on whole-genome sequencing (WGS), occurring between May 1, 2015, and Oct 31, 2018. A probable case was defined as laboratory-confirmed infection with S Enteritidis with the multiple-locus variable-number tandem repeat analysis outbreak profile. Multi-country epidemiological, trace-back, trace-forward, and environmental investigations were done. We did a case-control study including confirmed and probable cases and controls randomly sampled from the population registry (frequency matched by age, sex, and postal code). Odds ratios (ORs) for exposure rates between cases and controls were calculated with unmatched univariable and multivariable logistic regression. FINDINGS: 18 EU and EEA countries reported 838 confirmed and 371 probable cases. 509 (42%) cases were reported in 2016, after which the number of cases steadily increased. The case-control study results showed that cases more often ate in food establishments than did controls (OR 3·4 [95% CI 1·6-7·3]), but no specific food item was identified. Recipe-based food trace-back investigations among cases who ate in food establishments identified eggs from Poland as the vehicle of infection in October, 2016. Phylogenetic analysis identified two strains of S Enteritidis in human cases that were subsequently identified in salmonella-positive eggs and primary production premises in Poland, confirming the source of the outbreak. After control measures were implemented, the number of cases decreased, but increased again in March, 2017, and the increase continued into 2018. INTERPRETATION: This outbreak highlights the public health value of multi-country sharing of epidemiological, trace-back, and microbiological data. The re-emergence of cases suggests that outbreak strains have continued to enter the food chain, although changes in strain population dynamics and fewer cases indicate that control measures had some effect. Routine use of WGS in salmonella surveillance and outbreak response promises to identify and stop outbreaks in the future. FUNDING: European Centre for Disease Prevention and Control; Directorate General for Health and Food Safety, European Commission; and National Public Health and Food Safety Institutes of the authors' countries (see Acknowledgments for full list).

Seroprevalence of hantaviruses and Leptospira in muskrat and coypu trappers in the Netherlands, 2016.

Aims: Seoul orthohantavirus (SEOV) and Leptospira spp. are zoonotic pathogens with rats as main reservoir. Recently, the presence of SEOV in brown rats was reported in one region in the Netherlands. Brown rats are a frequent bycatch in traps placed to catch muskrats (Ondatra zibethicus) and coypus (Myocastor coypus), and thus are a potential health risk for trappers. It was our aim to determine the seroprevalence of orthohantavirus, specifically SEOV, and Leptospira spp in Dutch trappers. Methods and results: Participating trappers provided serum samples and completed an online questionnaire. The serum was tested for the presence of antibodies against six orthohantaviruses and eight Leptospira serovars. Two hundred-sixty trappers completed the online questionnaire (65%), and 246 (61%) and 162 (40%) serum samples were tested for relevant orthohantaviruses and Leptospira spp., respectively. The seroprevalence of Puumala orthohantavirus in Dutch trappers was 0.4% (95% CI: 0.1-2.3%). None of the participants tested positive for SEOV. The seroprevalence of leptospirosis was 1.2% (95% CI: 0.3-4.4%), although Leptospira spp. are present in brown rats in the Netherlands.Significance of study: The results indicate that the infections with orthohantaviruses and leptospires is low for muskrat and coypu trappers.

Diagnosis and Treatment of Listeria monocytogenes Endophthalmitis: A Systematic Review.

PURPOSE: Describe patient characteristics, treatment, and vision outcomes of Listeria monocytogenes endophthalmitis, an exceedingly rare form of listeriosis. METHODS: L. monocytogenes endophthalmitis cases in human adults, located through Medline (32) and from disease surveillance centers (11). L. monocytogenes conjunctivitis and keratitis were excluded. RESULTS: Most cases occurred in 2000-2015 (22/43), and almost all in Europe or North America (40/43). Patients were a median 61 years, 57% male (24/42) and half were immunosuppressed. Median days from entering care to diagnosis was 8 (IQR = 5-17). Only four were exogenous infections. L. monocytogenes was identified in 31/35 of anterior eye fluid samples (89%). Antibiotic regimens varied markedly (mostly ≥3 drugs). At diagnosis, most were blind in the affected eye (85%, 28/33), only a third regained normal vision (12/36). Older patients had poorer outcomes. CONCLUSIONS: Cases increased over time. Diagnostic delays were common and visual impairment often refractory to treatment, especially in older adults. The condition's rarity and variation in treatment makes it difficult to identify optimum therapy.

The economic burden of a Salmonella Thompson outbreak caused by smoked salmon in the Netherlands, 2012-2013.

Background: In 2012, the Netherlands experienced the most extensive food-related outbreak of Salmonella ever recorded. It was caused by smoked salmon contaminated with Salmonella Thompson during processing. In total, 1149 cases of salmonellosis were laboratory confirmed and reported to RIVM. Twenty percent of cases was hospitalised and four cases were reported to be fatal. The purpose of this study was to estimate total costs of the Salmonella Thompson outbreak. Methods: Data from a case-control study were used to estimate the cost-of-illness of reported cases (i.e. healthcare costs, patient costs and production losses). Outbreak control costs were estimated based on interviews with staff from health authorities. Using the Dutch foodborne disease burden and cost-of-illness model, we estimated the number of underestimated cases and the associated cost-of-illness. Results: The estimated number of cases, including reported and underestimated cases was 21 123. Adjusted for underestimation, the total cost-of-illness would be €6.8 million (95% CI €2.5-€16.7 million) with productivity losses being the main cost driver. Adding outbreak control costs, the total outbreak costs are estimated at €7.5 million. Conclusion: In the Netherlands, measures are taken to reduce salmonella concentrations in food, but detection of contamination during food processing remains difficult. As shown, Salmonella outbreaks have the potential for a relatively high disease and economic burden for society. Early warning and close cooperation between the industry, health authorities and laboratories is essential for rapid detection, control of outbreaks, and to reduce disease and economic burden.

International linkage of two food-borne hepatitis A clusters through traceback of mussels, the Netherlands, 2012.

This report describes an outbreak investigation starting with two closely related suspected food-borne clusters of Dutch hepatitis A cases, nine primary cases in total, with an unknown source in the Netherlands. The hepatitis A virus (HAV) genotype IA sequences of both clusters were highly similar (459/460 nt) and were not reported earlier. Food questionnaires and a case-control study revealed an association with consumption of mussels. Analysis of mussel supply chains identified the most likely production area. International enquiries led to identification of a cluster of patients near this production area with identical HAV sequences with onsets predating the first Dutch cluster of cases. The most likely source for this cluster was a case who returned from an endemic area in Central America, and a subsequent household cluster from which treated domestic sewage was discharged into the suspected mussel production area. Notably, mussels from this area were also consumed by a separate case in the United Kingdom sharing an identical strain with the second Dutch cluster. In conclusion, a small number of patients in a non-endemic area led to geographically dispersed hepatitis A outbreaks with food as vehicle. This link would have gone unnoticed without sequence analyses and international collaboration.

Extension of traditional infectious disease surveillance with a repeated population survey.

BACKGROUND: Public health surveillance is an important tool for monitoring cases of infectious diseases. Identification of risk factors requires the comparison of exposure between cases and controls. However, standard surveillance systems do not routinely collect information on controls. METHODS: Since 2008, we have extended the surveillance of infectious diseases in The Netherlands with a repeated population survey. This survey is based on the thrice-yearly administration of a questionnaire about potential risk factors for several gastrointestinal, foodborne and respiratory infections to a representative, dynamic sample of the Dutch population. The questionnaire contains questions similar to those used for interviewing cases. RESULTS: Over 14 mailing rounds, 4926 persons were approached with a response of 36%, with a small underrepresentation of men, young people, people living in large cities and persons with both parents born outside The Netherlands. Costs per completed questionnaire were around 15 euro. Muscle/joint pain in the past 4 weeks was the most reported symptom (44%), followed by running nose (39%) and headache (32%); 5.6% reported gastroenteritis, reflecting an incidence of 997 episodes per 1000 person-years. CONCLUSIONS: Extending traditional surveillance with a repeated population survey offers the unique opportunity to gather data for a multitude of purposes. The survey already has been used in two outbreak investigations and two case-control studies. It is cost-effective and may provide novel epidemiological insights towards risk group and risk factor identification and characterization for a variety of infectious diseases. The survey will be continued and expanded in use.

Cost-of-illness and disease burden of food-related pathogens in the Netherlands, 2011.

To inform risk management decisions on control and prevention of food-related disease, both the disease burden expressed in Disability Adjusted Life Years (DALY) and the cost-of-illness of food-related pathogens are estimated and presented. Disease burden of fourteen pathogens that can be transmitted by food, the environment, animals and humans was previously estimated by Havelaar et al. (2012). In this paper we complement these by cost-of-illness estimates. Together, these present a complete picture of the societal burden of food-related diseases. Using incidence estimates for 2011, community-acquired non-consulting cases, patients consulting their general practitioner, hospitalized patients and the incidence of sequelae and fatal cases, estimates were obtained for DALYs, direct healthcare costs (e.g. costs for doctor's fees, hospitalizations and medicines), direct non-healthcare costs (e.g. travel costs to and from the doctor), indirect non-healthcare costs (e.g. productivity loss, special education) and total costs. The updated disease burden for 2011 was equal to 13,940 DALY/year (undiscounted) or 12,650 DALY/year (discounted at 1.5%), and was of the same magnitude as previous estimates. At the population-level thermophilic Campylobacter spp., Toxoplasma gondii and rotavirus were associated with the highest disease burden. Perinatal listeriosis infection was associated with the highest DALY per symptomatic case. The total cost-of-illness in 2011 of fourteen food-related pathogens and associated sequelae was estimated at € 468 million/year, if undiscounted, and at € 416 million/year if discounted by 4%. Direct healthcare costs accounted for 24% of total costs, direct non-healthcare costs for 2% and indirect non-healthcare costs for 74% of total costs. At the population-level, norovirus had the highest total cost-of-illness in 2011 with € 106 million/year, followed by thermophilic Campylobacter spp. (€ 76 million/year) and rotavirus (€ 73 million/year). Cost-of-illness per infected case varied from € 150 for Clostridium perfringens intoxications to € 275,000 for perinatal listeriosis. Both incident cases and fatal cases are more strongly correlated with COI/year than with DALY/year. More than 40% of all cost-of-illness and DALYs can be attributed to food, in total € 168 million/year and 5,150 DALY/year for 2011. Beef, lamb, pork and poultry meat alone accounted for 39% of these costs. Products of animal origin accounted for € 86 million/year (or 51% of the costs attributed to food) and 3,320 DALY/year (or 64% of the disease burden attributed to food). Among the pathogens studied Staphylococcus aureus intoxications accounted for the highest share of costs attributed to food (€ 47.1 million/year), followed by Campylobacter spp. (€ 32.0 million/year) and norovirus (€ 17.7 million/year).

Comparison of five influenza surveillance systems during the 2009 pandemic and their association with media attention.

BACKGROUND: During the 2009 influenza pandemic period, routine surveillance of influenza-like-illness (ILI) was conducted in The Netherlands by a network of sentinel general practitioners (GPs). In addition during the pandemic period, four other ILI/influenza surveillance systems existed. For pandemic preparedness, we evaluated the performance of the sentinel system and the others to assess which of the four could be useful additions in the future. We also assessed whether performance of the five systems was influenced by media reports during the pandemic period. METHODS: The trends in ILI consultation rates reported by sentinel GPs from 20 April 2009 through 3 January 2010 were compared with trends in data from the other systems: ILI cases self-reported through the web-based Great Influenza Survey (GIS); influenza-related web searches through Google Flu Trends (GFT); patients admitted to hospital with laboratory-confirmed pandemic influenza, and detections of influenza virus by laboratories. In addition, correlations were determined between ILI consultation rates of the sentinel GPs and data from the four other systems. We also compared the trends of the five surveillance systems with trends in pandemic-related newspaper and television coverage and determined correlation coefficients with and without time lags. RESULTS: The four other systems showed similar trends and had strong correlations with the ILI consultation rates reported by sentinel GPs. The number of influenza virus detections was the only system to register a summer peak. Increases in the number of newspaper articles and television broadcasts did not precede increases in activity among the five surveillance systems. CONCLUSIONS: The sentinel general practice network should remain the basis of influenza surveillance, as it integrates epidemiological and virological information and was able to maintain stability and continuity under pandemic pressure. Hospital and virological data are important during a pandemic, tracking the severity, molecular and phenotypic characterization of the viruses and confirming whether ILI incidence is truly related to influenza virus infections. GIS showed that web-based, self-reported ILI can be a useful addition, especially if virological self-sampling is added and an epidemic threshold could be determined. GFT showed negligible added value.