Antimicrobial resistance and sickle cell protein analysis in invasive nontyphoidal salmonella isolated from outpatient and hospitalized children below 16 years in informal settlements in Nairobi, Kenya

Background: Invasive Non-typhoidal Salmonella (iNTS) disease continues to be a major public health problem, especially in sub Saharan Africa where incidence rates are 227 cases [range 152-341] per 100,000 population. Populations at risk of iNTS include adults with HIV infection, malnourished children, those with recent malaria or sickle-cell anaemia (SCA). Individuals with SCA are at an increased risk of invasive bacterial infections with the proportion of deaths from infection reported to be as high as 38% in the United States and 29% in Jamaica. Out of 2684 dry blood samples subjected to HPLC for investigation of sickle cell disease traits, 1820/2684 (67%) had normal hemoglobin (Hb AA/ Hb AF); (162/2684 (6%) tested positive for Sickle Cell Traits (Hb AS/Hb AFS); while 4/2684 (0.2%) tested positive for Sickle cell disease (Hb FS). The high MDR resistance phenotype in iNTS isolates and emerging resistance to third generation cephalosporins is of great concern in management of iNTS in our settings. Sickle cell disease was not a major factor among children with iNTS disease and no significant association with iNTS was observed. years, settlements,

In Kenya, iNTS disease is particularly a major challenge in poor informal settlements with infants and young children less than 5 years of age being the most affected; mortality rates can be 20-25% unless prompt treatment is administered.

Methods
Our study was conducted in 3 outpatient sites and 1 inpatient site, the outpatient sites were all located within Mukuru informal settlement, a densely populated slum, 15km East of Nairobi City. Blood and stool samples from children with fever alone and with fever and diarrhea were collected for processing for presence of iNTS using basic microbiology procedures. Dry blood spots were also taken and processed for sickle cell protein markers using High performance liquid chromatography (HPLC).

Results
A total of 22,246 blood and stool samples were collected from children < 16 years of age with fever/with or without diarrhea, for a period of 6 years and subjected to microbiological culture and detection of bacterial pathogens. Out of these 741 (3.3%) tested positive for Salmonella species. A total of 338/741(41%) NTS were isolated across all the sites; these consisted of 158/741(21%) Salmonella Enteritidis and 180/741 (24%) Salmonella Typhimurium.

Conclusion
The high MDR resistance phenotype in iNTS isolates and emerging resistance to third generation cephalosporins is of great concern in management of iNTS in our settings. Sickle cell disease was not a major factor among children with iNTS disease and no significant association with iNTS was observed.
In SSA, invasive non-typhoidal Salmonella (iNTS) disease is caused majorly by two serovars; Salmonella serovars Typhimurium and Enteritidis. It is a major challenge in SSA as it is responsible for increased cases of childhood morbidity and mortality. Serovar Salmonella Typhimurium is more common than Salmonella Enteritidis [6].
In most SSA countries, empirical treatment is often administered due to lack of strengthened laboratory systems in the health care facilities [7; 8]. This greatly contributes to inability to detect antimicrobial resistance which can persist in the patient's body system for many months [9,10].
Invasive NTS disease (iNTS) in children in SSA has been associated with malaria infection, malnourishment as well as sickle cell disease [11,12]. Besides iNTS infections, patients with sickle cell disease (SCD) are susceptible to a variety of other bacterial infections, which are a major cause of morbidity and mortality [13].
This increased susceptibility to infections is related to abnormalities in the defence mechanisms of these patients, including functional hypo-splenism, an abnormality in the alternative pathway of complement activities, and defective neutrophil function [13,14].
Devitalisation of the gut and bone due to repetitive vaso-occlusive crises, macrophage saturation with red cell breakdown products as a result of chronic haemolysis, and underlying splenic and hepatic dysfunction, are known to predispose patients with SCD to Salmonella infection [13]. Reduction of iNTS disease incidence has been reported following an improvement of effective malaria control measures [15].
Increasing antimicrobial resistance (AMR) in iNTS is of great global concern, and the situation is even more serious in low and middle-income countries where empiric treatment options for effective treatment of life-threatening invasive disease are limited. Several studies have shown that strains of NTS that are multidrug resistant to recommended first-line antibiotics, including ampicillin, Trimethoprim-sulfamethoxazole, chloramphenicol and kanamycin have emerged in several African countries over the past 20 years. Multi-drug resistance in iNTS has previously been reported in Kenya and Malawi (4,10,16) posing a major challenge to treatment and management options [17].
Salmonella Typhimurium, sequence type 313 (ST313) is a distinct phylogenetic lineage that has emerged in Africa. It is now a significant cause of iNTS disease in Africa. A recent study (18) reported isolating ST313 with both lineages I and II, but fewer ST19 strains, an important cause of iNTS disease as well as asymptomatic carriage.
Treatment failure and complications are associated with lack of proper diagnostic capacity that can aid in management of these multi-drug resistant strains [15,16].
We report on MDR iNTS from patients treated at outpatient clinics showing limited prevalence of sickle cell disease or sickle cell trait.

Study Site
We studied patients attending three outpatient facilities and one inpatient facility; These informal settlements are densely populated and characterized by limited basic services and infrastructure for providing clean water, sanitation facilities, solid-waste management, roads, drainage, and electricity [8]

Study Design
The study utilized a prospective longitudinal study design

Study Population
The study sought to recruit suspected cases of children < 16 years of age with bloodstream infections in Mukuru and Kibera informal settlement. We recruited patients presenting for care at 3 outpatient medical facilities and 1 inpatient facility that serves the informal settlements. These patients were approached for participation in the study if they were; residents in the Mukuru and Kibera slum; presented with a subjective history of at least 3 days of fever and have an axillary temperature of at least 37.5 o C or they presented with a history of fever of any duration and have an axillary temperature of at least 37.5 o C; or they reported having had three or more loose or liquid stools (children > 2 years) or 8 or more for infants in the 24 hours before presentation, or one or more loose or liquid stool with visible blood. We also included controls for our study, these were individuals (children) who presented themselves to the clinic to attend the mother child clinic.
The cases were therefore age matched with the controls. For every case, we sort for 2 controls. The controls ideally were healthy individuals recruited from our sites and residing in Mukuru or Kibera slums [8].

Blood for culture
For blood culture 1-3 ml for children < 5 years of age and 5-10 ml for 5 - and further typed by species-specific serological tests for Salmonella species [8].

Stool cultures
The rectal swab or loopful of the stool specimen was transported to KEMRI laboratory and initially cultured on selenite F (Oxoid, Basingstoke, UK) broth aerobically at 37 o C overnight. Broth cultures were then sub cultured on MacConkey agar and Salmonella-Shigella agar (Oxoid, UK) and incubated at 37 o C overnight. To identify suspect Salmonella bacteria, non-lactose fermenting colonies were biochemically tested using triple sugar iron (TSI) slants. From the subcultures, bacterial isolates were identified using biochemical tests on API20E strips and further typed by species-specific serological tests [8].

Malaria Test (Rapid Detection Test)
Using a sterile lancet, a gentle prick was made towards the pulp of the 4th finger at the disinfected site. The first drop of blood was expressed by applying gentle pressure to the finger and wiped away with a dry piece of cotton wool. Gentle pressure was applied to the same finger until a new blood drop appeared. Using the blood collection device provided in the RDT kit, the open end was gently immersed in the blood drop. The required volume of blood was collected as per manufacturer's instructions. The collected blood was transferred to the sample well. Holding the buffer bottle vertically, a drop of the buffer was added into the buffer well. Interpretation of the results was done according to the manufacturer's instructions.

High Performance Liquid Chromatography (HPLC)
Diagnosis of SCD was done using HPLC technique with samples run against a known range of SC protein standards. The HPLC was connected to the computer, which shows separation of the various haemoglobin types on a screen. HPLC offers the advantages such as full automation of the entire procedure and accurate quantification of the levels of haemoglobin present as long a variant or known Hb in the patient sample does not interfere with the interpretation [14,19].

Ethical Considerations
The study was approved by the Scientific and Ethics Review Unit (SERU) of KEMRI (SSC No. 2076). All parents and/or guardians of participating children were informed of the study objectives and voluntary written consent was sought and obtained before inclusion. A copy of the signed consent was filed and stored in password protected cabinets at KEMRI. All methods were carried out in accordance with relevant guidelines and regulations.

NTS AMR profiles from 2013-2018
Resistance trends for S. Typhimurium against the first line antibiotics were as follows: Ampicillin (

Occurrence of NTS and social demographic factors
A total of 338 (100%) were isolated. 180/338 (53.25%) were S. Typhimurium while 158/338 (46.74%) were S. Enteritidis. The P value was < 0.05 indicating that there was a significant association between the iNTS and the age groups isolated from.
Age group 0-5 years had most NTS isolated. Across all age groups females were more affected than males. It was also observed that stool samples had more NTS isolated as compared to blood samples Table 1.

Occurrence of NTS and Sickle cell disease/Trait
A total of 2684 samples from both blood and stool samples were subjected to a HPLC for processing of sickle cell disease. A total of 1820/2684(67.8%) tested normal haemoglobin, 162/2684(6.03%) were confirmed to have sickle cell trait while 4/2684(0.2%) tested positive for sickle cell disease. Interestingly 4/2684(0.2%) samples that tested positive for SCD tested negative for NTS in both blood and stool samples ( Table 2).

Occurrence of NTS in Malaria and HIV
A total of 8/338 (2.4%) children positive for NTS were also positive for Malaria but negative for HIV. These children came from a population of both cases and controls. S. Typhimurium was isolated in 7 malaria positive patients; 5 cases and 2 controls while S. Enteritidis was isolated in one malaria positive patient. A total of 5/338 (1.5%) patients positive for NTS tested positive for HIV. Among these 3 were cases while 2 were controls (healthy individuals) Table 4.  [6,20] hence causing a serious challenge in the management of infections especially in LMICs such as Kenya. These iNTS infections occur in urban informal settlements due to several factors such as crowding, street foods, poor drainage systems, poor hygiene and sanitation [8,21].
A previous study reported that the impact of iNTS on childhood mortality exceeds malaria in some African communities [24]. This study estimated mortality rates for iNTS among hospitalized patients in Africa to range from 4.4 to 27% for children [6;25] and 22 to 47% for adults [9,26].
The burden of disease due to iNTS is significant in sub Saharan Africa and South East Asia. [27]. It was reported that iNTS occurred in an estimated 88 cases per 100,000 person-years in the age group of 5 years old in rural Kenya, while in Mozambique, NTS was estimated to account for 120 cases per 100,000 personyears [27].
These incidences are likely to be under-estimates as many children with iNTS die before reaching the health facilities [25,28].
In SSA, risk factors for iNTS disease in children include HIV infection; malnutrition; [6] and malaria [29]. SCD is also an important risk factor in African children [30].
NTS bacteraemia overlaps significantly with malaria in Africa, both in terms of seasonality and affected age groups [31].
Many attempts have been made to explain how malaria causes susceptibility to NTS. The most consistent and likely evidence is that malarial haemolysis of the red blood cells creates favourable conditions for bacterial growth [32] This haemolysis prevents the effective eradication of NTS that successfully enter the systemic blood stream via the intestines. It remains inconclusive whether or not malaria infections facilitate the entry of NTS into the blood stream. Our study reported no significant association of iNTS with HIV, Malaria or even Sickle cell disease This is in contrast to studies done in the western part of Kenya where the malaria is endemic, and SCD and HIV are risk factors to iNTS disease [33].
The age group 0 to 5 years across all the sites had the highest number of iNTS infections A similar study in Kenya [34] revealed similar findings to our study that infants (below 1 year) were more likely to be infected with enteric pathogenic bacterial infection (OR 0.3, 95% CI 0.1-0.8) than the older ones. A recent study done in Lusaka Zambia also implicated a high prevalence of infection on the ages of 0-12 months (61.3%) [35].
Salmonella species were once susceptible to a broad range of affordable and effective antimicrobial drugs, but multidrug-resistant strains [9;36] have emerged in Africa. S Typhimurium was shown to contain a composite element encoding multidrug-resistant genes located on a virulence-associated plasmid (pSLT-BT), thus potentially linking resistance to antimicrobial drugs with virulence.
Resistance has necessitated the widespread use of expensive drugs for empirical management of sepsis, such as third generation cephalosporins and fluoroquinolones (eg, ciprofloxacin), which are majorly unaffordable in our settings, and which could promote the development of further resistance [36,37].

Conclusion
It is evident that public health education on improved WASH techniques is essential to control sanitation-related infections. However, as WASH infrastructure will be long in coming especially in overcrowded living conditions in informal settlements, introduction of a bivalent efficacious vaccine that would target Salmonella serovars Typhimurium and Enteritidis would significantly lower the disease burden in high-risk populations.