Changing epidemiology of invasive non-typhoid Salmonella infection: a nationwide population-based registry study

Abstract Objectives Non-typhoid Salmonella (NTS) may invade beyond the intestine, causing bacteraemia, sepsis, and infection of normally sterile sites. The epidemiology of invasive NTS (iNTS) infection is under-researched. We determined trends, risk factors, serotype distribution, antimicrobial resistance (AMR), and attributable sources of iNTS infection in a high-income setting. Methods 22,837 records of culture-confirmed human salmonellosis cases and 10,008 serotyped Salmonella isolates from five putative animal reservoirs (pigs, cattle, broilers, layers, reptiles) in the Netherlands during 2005-2018 were retrieved from national surveillance registries. Risk factors for iNTS infection were identified using logistic regression analysis. Source attribution modelling was based on serotyping, prevalence, and exposure data. Results The average annual percentage of iNTS infections was 4.6% (range: 3.5-5.7%). An increase in iNTS infections was observed since 2012 (Odds Ratio [OR]: 1.09, 95% Confidence Interval [95%CI]: 1.04-1.14). Increased iNTS infection risk was associated with wintertime (OR: 1.37, 95%CI: 1.12-1.66), male sex (OR: 1.73, 95%CI: 1.51-1.99), older age (ORs: 3.27 to 16.33, depending on age groups), and living in rural areas (OR: 1.54, 95%CI: 1.23-1.93). While 52% of iNTS infections (n=950) were caused by serotypes Enteritidis and Typhimurium, those displaying the highest invasiveness relative to their occurrence were Dublin (32.9%, n=163), Panama (21.6%, n=106), and Poona (14.1%, n=71). Cattle were a larger source of iNTS than non-iNTS infections (12.2% vs. 7.6%). Lower AMR and multi-resistance rates were observed among iNTS (37.9%) than non-iNTS isolates (48.6%). Conclusions The increase in iNTS infections, which is reported also in other countries, is of public health and clinical concern. The underlying reasons seem to be multi-factorial in nature. iNTS infection risk depends more on the infecting serotypes and patient demographics, and less on the attributable reservoirs and AMR profiles.