Case-control investigation of invasive Salmonella disease in Malawi reveals no evidence of environmental or animal transmission of invasive strains, and supports human to human transmission

Abstract

Background: Invasive Salmonella infections cause significant morbidity and mortality in Sub-Saharan Africa. However, the routes of transmission are uncertain. We conducted a case-control study of index-case and geographically-matched control households in Blantyre, Malawi, sampling Salmonella isolates from index cases, healthy people, animals, and the household environment.

Methodology: Sixty index cases of human invasive Salmonella infection were recruited (March 2015-Oct 2016). Twenty-eight invasive Non-Typhoidal Salmonella (iNTS) disease and 32 typhoid patients consented to household sampling. Each index-case household was geographically matched to a control household. Extensive microbiological sampling included stool sampling from healthy household members, stool or rectal swabs from household-associated animals and boot-sock sampling of the household environment.

Findings: 1203 samples from 120 households, yielded 43 non-Typhoidal Salmonella (NTS) isolates from 25 households (overall sample positivity 3.6%). In the 28 iNTS patients, disease was caused by 3 STs of Salmonella Typhimurium, mainly ST313. In contrast, the isolates from households spanned 15 sequence types (STs). Two S. Typhimurium isolates from index cases closely matched isolates from their respective asymptomatic household members (2 and 3 SNP differences respectively). Despite the recovery of a diverse range of NTS, there was no overlap between the STs causing iNTS disease with any environmental or animal isolates.

Conclusions: The finding of NTS strains from index cases that matched household members, coupled with lack of related animal or environmental isolates, supports a hypothesis of human to human transmission of iNTS infections in the household. The breadth of NTS strains found in animals and the household environment demonstrated the robustness of NTS sampling and culture methodology, and suggests a diverse ecology of Salmonella in this setting. Healthy typhoid (S. Typhi) carrier state was not detected. The lack of S. Typhi isolates from the household environment suggests that further methodological development is needed to culture S. Typhi from the environment.

Fig 1

The geographic location of samples in this study a) & b) are for typhoid fever case and control households, while c) and d) are for iNTS case and control households. a) and c) indicate the number of households (case + control) that were sampled from each area, in parentheses is the number of households where Salmonella was isolated (not including the index case isolate). b) and d) indicated the sequence type of all the Salmonella isolated from that location, including the index case isolates. STs are coloured by the type of sample they were isolated from. Red is invasive Salmonella, blue is a household member of an invasive Salmonella case, purple is in both invasive Salmonella and a household member, brown is the environment and orange is animal. Map shapefile obtained from https://data.humdata.org/dataset/20eb8e5b-134d-41d8-a56f-4f358f7faf16/resource/50f185b1-b028-4787-a591-80c8db81cfed/download/mwi_adm_nso_20181016_shp.zip. The map shape files are licenced under the Creative Commons Attribution for Intergovernmental Organisations licence - https://creativecommons.org/licenses/by/3.0/igo/legalcode.

Fig 2. Minimum spanning tree based on seven MLST genes showing the genetic similarity between different niches–the only non-clinical niche where STs associated with human invasive disease were present was the gut of household contacts.

Each circle represents a sequence type, the size of the circle is proportional to the number of isolates of that ST. STs that vary by one locus are joined by a black bar. Two invasive disease index isolates had a matched isolate from a healthy human sample of the same ST from the same household. These isolates are indicated by black and red outlines. In cases where there were multiple isolates of the same sequence type isolated from the same household, these “slices” of pie have had a border of the same colour. I.e. There were two isolates of ST3257 isolated from the same household, one from human invasive disease, and one from human nont invasive disease. Therefore, on the ST3257 pie, one slice of orange (not-invasive) and one slice of light blue (invasive) have been bordered in red.

Fig 3. The close relationship between the index case and family member isolates from households 25 and 44 shown with a phylogenetic tree.

Whole genome maximum likelihood tree of 19 S. Typhimurium isolates described in Table 2. The numbers at the end of each branch indicate household, as described in the first column of Table 2 (all these households are case households). Sequencing read sets were compared against the reference genome D23580 as described in Methods.