Nontyphoidal Salmonellae commonly cause invasive disease in African children that is often fatal. The clinical diagnosis of these infections is hampered by the absence of a clear clinical syndrome. Drug resistance means that empirical antibiotic therapy is often ineffective and currently no vaccine is available. The study objective was to identify risk factors for mortality among children presenting to hospital with invasive Salmonella disease in Africa. We conducted a prospective study enrolling consecutive children with microbiologically-confirmed invasive Salmonella disease admitted to Queen Elizabeth Central Hospital, Blantyre, in 2006. Data on clinical presentation, co-morbidities and outcome were used to identify children at risk of inpatient mortality through logistic-regression modeling. Over one calendar year, 263 consecutive children presented with invasive Salmonella disease. Median age was 16 months (range 0–15 years) and 52/256 children (20%; 95%CI 15–25%) died. Nontyphoidal serovars caused 248/263 (94%) of cases. 211/259 (81%) of isolates were multi-drug resistant. 251/263 children presented with bacteremia, 6 with meningitis and 6 with both. Respiratory symptoms were present in 184/240 (77%; 95%CI 71–82%), 123/240 (51%; 95%CI 45–58%) had gastrointestinal symptoms and 101/240 (42%; 95%CI 36–49%) had an overlapping clinical syndrome. Presentation at <7 months (OR 10.0; 95%CI 2.8–35.1), dyspnea (OR 4.2; 95%CI 1.5–12.0) and HIV infection (OR 3.3; 95%CI 1.1–10.2) were independent risk factors for inpatient mortality. Invasive Salmonella disease in Malawi is characterized by high mortality and prevalence of multi-drug resistant isolates, along with non-specific presentation. Young infants, children with dyspnea and HIV-infected children bear a disproportionate burden of the Salmonella-associated mortality in Malawi. Strategies to improve prevention, diagnosis and management of invasive Salmonella disease should be targeted at these children.
Nontyphoidal Salmonellae are a major, yet often neglected, cause of fatal invasive disease among young African children, responsible for over 100,000 deaths a year. There is currently a lack of prospective studies to understand the clinical presentation and course of this disease. Diagnosis can be confusing, antibiotic resistance is an increasing problem and no vaccine is available. Studying 263 consecutive cases of invasive Salmonella disease over a one year period, we observed disparate clinical presentations, precluding the possibility of making a reliable clinical diagnosis in cases of this disease. We confirmed high mortality rates, finding that young age, dyspnea and HIV infection are independent risk factors for death. These findings support both the need for improved measures for prevention, diagnosis and management of invasive nontyphoidal Salmonella disease and indicate which children with this disease are at most need of intensive clinical care.
Citation: MacLennan CA, Msefula CL, Gondwe EN, Gilchrist JJ, Pensulo P, Mandala WL, et al. (2017) Presentation of life-threatening invasive nontyphoidal Salmonella disease in Malawian children: A prospective observational study. PLoS Negl Trop Dis 11(12): e0006027. https://doi.org/10.1371/journal.pntd.0006027
Editor: Nicholas Feasey, Liverpool School of Tropical Medicine, UNITED KINGDOM
Received: May 21, 2017; Accepted: October 10, 2017; Published: December 7, 2017
Copyright: © 2017 MacLennan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by a Wellcome Trust Research Fellowship [grant number 067902/Z/02/Z] to CAM, Wellcome Trust PhD studentships to CLM and ENG, a Wellcome Trust Programme Grant [grant number 074124/Z/ 04/Z] to MEM., a PhD studentship from the Gates Malaria Partnership (to WLM) which received support from the Bill and Melinda Gates Foundation, and a Clinical Research Fellowship from GlaxoSmithKline to CAM. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: CAM was the recipient of a clinical research fellowship from GlaxoSmithKline. This does not alter our adherence to all PLOS policies on sharing data and materials.
Invasive bacterial disease is a major cause of mortality among African children [1–4]. Prospective studies from rural-based district hospitals in Kenya and Mozambique have found community-acquired bacteremia to be responsible for 26%  and 21%  of pediatric inpatient deaths. Pneumococcus and nontyphoidal Salmonellae (NTS) are the most common causes of bacteremia among African children [1–5]. While the burden of disease due to pneumococcus should fall with implementation of pneumococcal conjugate vaccines, no vaccine is available or in clinical trials against invasive nontyphoidal Salmonella (iNTS) disease [6, 7].
iNTS disease among African children is caused primarily by Salmonella enterica serovars Typhimurium and Enteritidis [8–12], and has an associated case fatality of around 20–25% for bacteremia [8, 9, 11, 13, 14] and 52% for meningitis . These findings contrast with Salmonella disease in industrialized nations which typically presents as self-limiting enterocolitis, rarely requiring hospitalization . Despite reports of declining iNTS disease with falling prevalence of malaria from specific African sites [17–19], recently published data from multiple countries across sub-Saharan Africa indicate that NTS continues to be a major cause of childhood bacteremia in west [20,21], central , east [23,24] and southern Africa [25,26]. The Phase 3 RTS,S/AS01 malaria vaccine trial conducted in eleven sites across sub-Saharan Africa, agnostic to levels of Salmonella bacteremia, found an incidence of approximately 500 cases of Salmonella sepsis/100,000 children/year among children under two years . Recent data from the Typhoid Surveillance in Africa Program across a further seven sub-Saharan African sites confirm that iNTS disease continues to be a major public health problem in the region .
The clinical presentation of iNTS disease is often a non-specific febrile illness [7–10, 14]. Further diagnostic challenges result from the association of iNTS disease with other co-morbidities [7–10, 14]. Among African children, iNTS disease often occurs with malaria [29–31], HIV [26, 31], malnutrition  and anemia . In adults, there is a strong association with HIV infection [26, 34]. Increasing multi-drug resistance among iNTS isolates in Africa adds to the high attrition from iNTS disease [8, 10–12]. For Malawian isolates, 90% have been reported as resistant to ampicillin, chloramphenicol and cotrimoxazole .
We set out to provide data to guide improved management of children with iNTS disease. Consecutive microbiologically-confirmed cases admitted to a large government hospital in Malawi were studied for the relationship between clinical presentation and outcome.
Queen Elizabeth Central Hospital (QECH) is the largest hospital in Malawi. It is situated in Blantyre, one of the two main cities in Malawi, and serves a combined urban and peri-urban rural population of around 1 million. The hospital has approximately 1000 beds and admits around 10,000 adults and 30,000 children (<16 years) annually. Since 1996, the Malawi-Liverpool-Wellcome Trust Clinical Research Programme has performed blood and cerebrospinal fluid (CSF) cultures from patients admitted with suspected sepsis and meningitis.
Study patients and clinical methods
Eligible participants were children age ≤ 14 years admitted to the Pediatric Department of QECH with isolation of Salmonella from blood and/or CSF between January 1st and December 31st 2006. Subjects not meeting all three eligibility criteria were excluded. In order to reduce bias due to early death following admission, children were approached for recruitment on the day that Gram-negative bacteria were first detected in the blood or CSF, either on initial Gram stain or following positive culture.
First-line antimicrobial therapy for suspected sepsis in children at the time of the study was intramuscular chloramphenicol and gentamicin, or penicillin and gentamicin. The great majority of Gram-negative bacteraemias detected were due to NTS, and these bacteria were almost all resistant to chloramphenicol and penicillin. Therefore, children were started on oral ciprofloxacin and/or intravenous ceftriaxone as soon as Gram-negative bacteria were first detected. Once the organism and antibiotic sensitivity profile were known, treatment would be amended as necessary.
First line antimicrobial therapy for pediatric meningitis was intravenous ceftriaxone. Demographic and clinical data collected on admission were recorded on standard forms, along with weight and height or length. Clinical progress was recorded daily through to discharge from hospital or inpatient death.
Respiratory distress was defined as the presence of a tracheal tug, intercostal or subcostal recession, head-bobbing or nasal flaring. Tachypnea and tachycardia on admission were defined as rates greater than the 90th centile for age . Children with an admission axillary temperature >37.5°C were classified as febrile. Children were considered to have gastroenteritis if diarrhea (≥3 loose stools per day) or vomiting were present, and respiratory disease if there was shortness of breath or respiratory distress. Fever without focus was defined as fever in the absence of localizing clinical signs.
Severe malnutrition in children 6–60 months was defined as a weight-for-height greater than 3 z-scores below the WHO median, or bilateral pedal edema (kwashiorkor). In children over 60 months, severe malnutrition was defined as a body mass index (BMI)-for-age greater than 3 z-scores below the WHO median. Severe anemia was defined as a packed cell volume (PCV) <15% or a hemoglobin concentration <5.0 g/dl. Hypoglycemia was defined as a glucose concentration on admission <3.0 mmol/L.
Sampling and laboratory methods
Blood and CSF cultures were performed using a standard aerobic bottle (BacT/Alert, bioMérieux, France) using 1–2 ml of blood and 0.5–1 ml of CSF. Following detection of Gram-negative bacteria, Salmonellae were confirmed by biochemical testing with API 20E kits (bioMérieux) and serovar was determined using agglutinating antisera (Prolab Diagnostics). Antimicrobial susceptibility testing was performed by disc diffusion using ampicillin, chloramphenicol, cotrimoxazole, gentamicin, ceftriaxone and ciprofloxacin discs according to British Society of Antimicrobial Chemotherapy methods and breakpoints. The laboratory participates in the UK National External Quality Assurance Scheme. All children admitted to QECH have a malaria parasite slide and PCV performed. Full blood counts were determined using a HMX (Becton Coulter). Tests for HIV infection were with Determine (Abbot Laboratories) and UniGold (Trinity Biotech) rapid tests, following national Voluntary Counseling and Testing guidelines. For discordant results and children under 18 months with positive results, the diagnosis was confirmed by polymerase chain reaction for proviral DNA.
Monthly rainfall data through 2006 were collected from the meteorological station at Chileka Airport, Blantyre.
Data were entered into a Microsoft Access database and analysis was performed using R (http://www.R-project.org/). Variation with age and season, of clinical indices, co-morbidity and outcome, were calculated as Mantel-Haenszel odds ratios. To facilitate this, age was converted into four ordinal categories (0–6 months, 7–12 months, 1–2 years, >2 years). Presentation season was grouped as January-March, April-June, July-September and October-December and ranked by mean rainfall. Odds ratios of case fatality given the presence of clinical indices or co-morbidity adjusted for age and season of presentation were calculated by logistic regression. Finally, a model of case fatality was constructed and tested by multivariate logistic regression. Covariates were chosen for the first fit of the model if they predicted mortality with a p value <0.1 along with age and sex. To minimize the number of covariate patterns, age was collapsed into a binary variable (children 0–6 months and children 7 months and over) for the purposes of the model. Covariates with the largest remaining p value (but retaining age and sex) were sequentially removed from the model and the model re-evaluated for goodness-of-fit and distribution of residuals at each step. The model was also evaluated for two-way interactions between significantly-associated terms. Missing data were addressed by providing denominator information where applicable.
After ensuring that each child was on appropriate antimicrobial therapy, the study was explained to the parents or guardians and informed written consent obtained. Ethical approval for the study was granted by the College of Medicine Research and Ethics Committee, College of Medicine, University of Malawi.
In total, 4535 blood cultures and 1823 CSF cultures were performed on children admitted to QECH over one year in 2006. Gram-negative organisms were detected in 395/4535 (9%) blood cultures and 34/1823 (2%) CSF cultures. Of these, 257/395 (65%) and 12/34 (35%), respectively, were identified as Salmonellae (Fig 1A). Salmonellae were isolated from both blood and CSF of six children, giving a total of 263 episodes of invasive Salmonella disease. The majority of Salmonellae 234/263 (89%) were serovar Typhimurium, while 14/263 (5%) were serovar Enteritidis and only 8/263 (3%) Typhi. 7/263 (3%) could not be typed with locally available antisera. Nine children had either a single (n = 5) or double (n = 4) recurrence of NTS bacteremia. NTS bacteraemia occurred in two sets of twins.
(A) Blood and CSF cultures taken from children admitted to Queen Elizabeth Central Hospital, Blantyre, Malawi in 2006; those yielding Salmonella isolates; and outcomes of children with invasive Salmonella disease. (B) Recruitment pathway for children following detection of Gram-negative bacteria in blood and/or CSF culture. (C) Kaplan-Meier estimate of survival (in days) of children following admission with invasive Salmonella disease.
Of the 263 children with microbiologically-confirmed invasive Salmonella disease, 35 (13%) had died and 23 (9%) been discharged prior to recruitment, leaving 205 (78%) whose parents and guardians were approached for consent. Consent was obtained for 178 (87%) out of the 205 children seen (Fig 1B). Median age of children was 16 months (range 0 months to 15 years) with 213/263 (81%) under 3 years (Table 1 and Fig 2A). The incidence of disease was associated significantly with rainfall (linear regression coefficient 4.70, 95% CI 2.48–6.92, p = 0.001), with 152/263 (58%) presenting between January and March during the rainy season (which last from November to April), and 30/263 from June to September in the dry season (11%) (Fig 2B). 7/263 (3%) children self-discharged or were not admitted.
Graphs indicate variation in mortality (A and B), co-morbidity (C and D) and clinical presentation (E and F) with age and month of the year. Each plot is overlaid on a histogram of the frequency of iNTS disease presentation with age (A, C and E) or bar-charts indicating monthly frequency of iNTS disease and monthly rainfall (B, D and F).
Clinical features and comorbidities of children with invasive Salmonella disease are shown in Table 2 and Fig 2. Most children presented with a history of fever and were febrile on admission. Cough was significantly more common as a presenting symptom than either diarrhea (p<0.001) or vomiting (p<0.001), with signs of respiratory distress in 72/166 (43%) of children at admission. Older children (≥7 months) tended to present more often with fever in isolation than younger children (<7 months; OR = 1.26, 95% CI 1.02–2.20). Neither the proportions of respiratory nor gastrointestinal presentations varied with age (Fig 2E), and clinical presentation did not vary significantly with season of presentation (Fig 2F).
45/180 (25%) of patients were severely malnourished. Median PCV on admission was 25%, and 47/244 (19%) of children were severely anemic. 38/249 (15%) had concurrent Plasmodium falciparum parasitemia. Of children tested, 70/162 (43%) were HIV-infected. Significantly more older children (≥7 months) were HIV-infected than younger children (<7 months, OR = 1.55; 95% CI 1.15–2.09). Malaria and malnutrition did not vary significantly with age (Fig 2C). Children diagnosed with iNTS disease during the months with high rainfall were less likely to be HIV-infected than those admitted in the dry season (OR = 0.67; 95% CI 0.49–0.89). The rates of malaria and malnutrition did not vary significantly with season (Fig 2D).
Antimicrobial resistance and usage
Most strains were resistant to ampicillin, cotrimoxazole and chloramphenicol (multi-drug resistant; 211/259, 81.5%). None were resistant to ciprofloxacin or ceftriaxone. Only 15/244 (6%) of children received first-line antimicrobials on admission to which their Salmonella isolate was susceptible in vitro. Conversion to either ceftriaxone or ciprofloxacin occurred with a median delay of 3 days (range 1–12 days).
Inpatient mortality was 52/256 (20%) (Fig 1C), highest among infants of 0 to 6 months inclusive (53%) (Fig 2A), and declined with increasing age (OR = 0.56, 95%CI 0.42–0.75; age in this model was divided into four ordinal categories, children 0–6 months, 7–12 years, 13–24 months, and >24 months). Mortality did not vary significantly with season (Fig 2B). Death occurred on the calendar day of admission in 8/52 (15%) of children who died, with a median time from admission to death of 3 days (range 0–18 days). Adjusting for age and season, significant risk factors for mortality were: a history of dyspnea, absence of fever, presence of respiratory distress or hypoglycemia at presentation, and HIV infection (Table 2). Cotrimoxazole preventive therapy (CPT) was not associated with mortality in HIV-infected children (OR = 2.17, 95%CI = 0.54–10.88; p = 0.30). 147/204 (72%) of children surviving their acute admission were followed up four to six weeks post-discharge from hospital (median time to follow up: 43 days), with no reported deaths in the intervening time. Only one out of the nine children with recurrent NTS bacteremia died.
A logistic regression model predicting mortality including age, sex, HIV status and a history of dyspnea was shown to be statistically significant (likelihood ratio test χ2 = 21.69, p = 0.0002) with a pseudo-R2 of 0.179. The Hosmer-Lemeshow goodness-of-fit test for the model provides no evidence to reject the model (p = 0.73) and the distribution and influence of the residuals (S1 Fig) demonstrate that the model is consistent with the data. In this model, age less than 7 months (OR = 10.0; 95% CI 2.8–35.1), HIV infection (OR = 3.3, 95% CI 1.1–10.2) and a history of dyspnea (OR = 4.2; 95% CI 1.5–12.0) were significant independent risk factors for death. Inpatient survival in children with and without each of these risk factors for mortality is presented in Fig 3.
Kaplan-Meier estimates of survival (in days) following admission with invasive Salmonella disease in children aged 0–6 months and >6 months (A), children with and without HIV co-infection (B), and children with and without a history of dyspnea at presentation (C).
Salmonellae are a common cause of invasive disease in African children that is often fatal. The clinical features and high mortality in the current study are unchanged from a retrospective study from the same hospital a decade earlier . Almost all patients had nontyphoidal serovars, most commonly Typhimurium, and the majority of isolates were MDR. In some settings, particularly urban centers, typhoid fever is more common than iNTS disease [12,24].
Invasive Salmonella disease occurred mainly in children under three years of age and, as previously reported, was more common in the rainy season [8, 9, 21, 35]. In 206/240 (86%) of children, the disease presented with fever together with respiratory symptoms, gastrointestinal symptoms or both of these (Fig 4A). Since a clinical diagnosis of invasive Salmonella disease is not reliable, laboratory detection of Salmonellae from blood or CSF by microbiological culture is required. Facilities for this are uncommon in African hospitals and, even where present, a diagnosis usually takes a minimum of 48 hours. Despite NTS being primarily considered an enteric pathogen, the majority of children presented with cough or shortness of breath, commonly in the absence of enteric symptoms, as previously observed [8–10, 38].
(A) Area-proportionate Venn diagram indicating the different presenting clinical syndromes of pediatric invasive Salmonella disease. Gastroenteritis is defined as children with a history of diarrhea or vomiting, and a respiratory presentation as the presence of cough, shortness of breath or respiratory distress. (B) Area-proportionate Venn diagram indicating comorbidities (HIV, malaria and severe malnutrition) of children presenting with invasive Salmonella disease. Figures are absolute numbers (percentages in brackets) of children in each group, with mortality in each group given below and are shown for individual variables and for children in whom these overlapped. Discrepancies between totals in the figures and the text are due to children with missing data for one of the categories depicted.
This partly explains why the large majority of children with iNTS disease did not receive empirical antimicrobials with good anti-Salmonella activity on admission. Conversion to an antimicrobial regimen with adequate Salmonella cover occurred with a median delay of 3 days from presentation, by which time half of the observed deaths had occurred. Therefore, in settings where NTS are responsible for a large proportion of invasive pediatric bacterial infections, empirical antimicrobial therapy for suspected bacteremia and sepsis should take into account local antimicrobial sensitivity of NTS. There has been a decline in antimicrobial resistance among iNTS isolates in Malawi since this study took place, but the majority of S. Typhimurium were still multidrug resistance in 2014. A small number of Salmonella isolates with altered susceptibility to ciprofloxacin and 3rd-generation cephalosporins have been reported in recent years .
The rapid demise of Malawian children with iNTS disease is similar to findings from Kenya . This is a major challenge for the management of children presenting with a syndrome indistinguishable from severe pneumonia, for which WHO guidelines recommend the use of ampicillin and chloramphenicol . Early deaths also increase the difficulty of studying iNTS disease, as children will often die before a microbiological diagnosis of Gram-negative bacteremia or meningitis is possible. This potential confounder can only be overcome by recruiting every child at the time of admission. With up to 400 pediatric admissions a day at QECH, this was not a feasible option in our setting.
Even in clinical facilities with diagnostic laboratories and appropriate antimicrobial therapy, the burden of iNTS disease remains high. Until a vaccine against iNTS disease becomes available and widespread immunization of children across sub-Saharan Africa is implemented, development of a rapid diagnostic that can identify children with iNTS disease in a timely manner should remain a public health priority. First-generation tests would be unlikely to provide information on antimicrobial resistance. Multidrug resistance (defined as resistance to ampicillin, chloramphenicol and cotrimoxazole) is common among NTS blood culture isolates across Africa [40, 41]. Therefore, in countries where conventional microbiological surveillance and antimicrobial resistance testing are lacking, treatment of NTS in the absence of information on antimicrobial resistance should avoid these antibiotics.
The possible emergence of new antimicrobial resistance patterns underlies the ongoing need for blood culture surveillance data with antimicrobial resistance testing. Owing to the ability of Salmonella to sequester in the intracellular niche within macrophages and cause latent/recrudescent infection, antibiotics with good intracellular penetration, such as ciprofloxacin, are advisable. The lack of reported deaths in the four to six week follow-up period emphasizes the importance and value of providing intensive clinical care to children with iNTS disease through their acute admission.
Comorbidity was common, with 86/139 (62%) of children presenting with at least one of: P. falciparum parasitemia, severe malnutrition or HIV infection (Fig 4B). Only 38/249 (15%) had concurrent parasitemia, fewer than has been reported previously [8, 9]. In contrast to previous reports , this proportion did not vary significantly with season, suggesting other reasons for seasonal fluctuation in iNTS disease cases, such as worsening sanitation during the rainy season facilitating transmission. We did not test for recent malaria, e.g. by detection of HRP2 in plasma. Recent malaria can be more strongly associated with iNTS disease in children than current malaria . Hemolysis-mediated immunological defects , which predispose children with malaria to iNTS disease, persist for several weeks following parasite clearance .
The dominant co-morbidity was HIV infection, affecting 70/162 (43%) of the children in the study. This greater proportion than has been reported previously [8, 9], likely reflects the systematic testing for HIV infection in the study. In other studies, HIV testing has often only been performed on a subset of children where HIV/AIDS has been suspected clinically (for example ). The significance of HIV infection as a risk factor for iNTS disease may vary across Africa, being more important in Southern and Eastern Africa compared with Western Africa owing to regional differences in HIV prevalence.
Background prevalence of HIV infection among children in Malawi at the time of the study was not known. However, available data of antenatal HIV prevalence and the numbers of children living with HIV in Malawi at the time of the study, suggest that the prevalence for the group under-five years was less than 2% . Data on HIV status were only available for 161/263 (61%) of the children, so these findings should be viewed with some caution. Nevertheless, proportionately more children who died lacked HIV data (29/52, 56%), as compared to those who survived (66/204, 32%; p = 0.003 for comparison of proportions), and so the association of HIV with mortality may underestimate the prevalence of HIV co-infection.
This study aimed to identify the children in whom iNTS disease is most likely to be fatal. Children with HIV infection, those with a history of dyspnea, and children under 7 months of age are all at significantly increased risk of dying. The odds of dying with iNTS disease are 10 times greater in children under 7 months than they are in older children. This substantially-increased risk of mortality in young infants is particularly striking (Fig 3A) and is observed in spite of a significantly lower prevalence of HIV infection in this age group. Immune naivety is the most likely reason. Infants are at particular risk of iNTS disease after maternally-acquired antibody is lost and before they produce their own NTS-specific antibodies [45–47].
The dyspnea associated with increased risk of dying is most likely due to metabolic acidosis resulting from Salmonella sepsis, or to bona fide Salmonella respiratory tract infection. Arterial blood gas and anion gap measurements were not available, and chest radiographs not commonly performed in these children. Hence, it was not possible to distinguish the underlying pathological mechanism. A history of shortness of breath often results in a child being misclassified as suffering from pneumonia in resource-poor healthcare settings , increasing the chance of children in this high risk subgroup receiving inappropriate empirical antimicrobials.
Although the association between HIV and iNTS disease is well established, the underlying mechanisms of HIV infection as a determinant of mortality in children with iNTS disease has not been well explored. Immunological defects in HIV-infected adults leading to susceptibility to iNTS disease include impaired gut mucosal immunity , and dysregulated cellular  and humoral immunity [50, 51]. The mechanisms that contribute to increased mortality in children with iNTS disease and HIV provide important questions for future research, which should explore the key interaction of HIV infection and malnutrition .
Only three children at presentation were receiving anti-retroviral therapy (ART), while 37 were on CPT. CPT has been shown to protect against iNTS disease in HIV-infected African adults prior to the use of ART  and subsequent studies have shown benefit of CPT in protecting against invasive bacterial disease in communities with high levels of in vitro cotrimoxazole resistance . However, CPT did not affect outcome in our study among HIV-infected children with iNTS disease. Recent work has indicated an important role for the introduction of ART among HIV-infected African children in reducing the incidence of iNTS disease [26, 31], but has yet to show that ART improves outcome among HIV-infected children with iNTS disease .
The main limitation of the study is that these data were collected from one site during one period in time, the calendar year 2006. New prospective studies from other sites in sub-Saharan Africa would be valuable for confirming the reproducibility of our findings across the region and their ongoing applicability to iNTS disease in Africa. In Blantyre, the epidemiology of invasive Salmonella disease changed markedly in 2011 with the start of an epidemic of typhoid fever. By 2012 S. Typhi was a more common blood culture isolate than NTS , indicating the fluidity of the epidemiology of invasive Salmonella disease in one site.
Additionally, iNTS disease has declined in Blantyre, accompanied by reductions in malaria, HIV (following the successful implementation of an antiretroviral therapy program beginning in 2005) and acute malnutrition . Nevertheless, across sub-Saharan Africa, iNTS disease remains the commonest presentation of invasive Salmonella disease and a major cause of pediatric bacteremia [2,8,27].
This prospective study provides additional evidence that iNTS disease is common and frequently fatal among children in sub-Saharan Africa, and particularly highlights the major diagnostic and management challenges associated with iNTS disease . iNTS disease does not present with a readily-identifiable clinical syndrome and current empirical antimicrobials do not provide effective treatment for Salmonella infections. Empirical antimicrobial treatment for potential Salmonella infection should be considered for children presenting with a clinical syndrome compatible with systemic infection. This is especially true for children with HIV infection and young infants, and particularly applies to presentations suggestive of community-acquired pneumonia in regions where invasive Salmonella disease is common.
The development of cheap rapid diagnostic tests for iNTS disease in children should be a priority. The poor effectiveness of antimicrobials means that other approaches to treatment are required. Epidemiological studies provide strong evidence that antibodies bactericidal to iNTS in the presence of complement protect against these infections [45, 46]. The development and use of a vaccine that induces protective antibodies in young infants may provide effective protection [6, 7].
S1 Fig. Distribution and influence of residuals in logistic regression model of mortality.
The graph demonstrates for each covariate pattern the squared standardized Pearson residual (on the y axis) and the probability of mortality for an individual with that covariate pattern (x axis). The area of each data point is proportional to influence statistic. The graph demonstrates no relationship between the probability of mortality for a given covariate pattern and the magnitude of the residuals. Residual analysis of the 14 covariate patterns demonstrates a single covariate pattern with a squared Pearson standardized residual greater than 3.84 (95% significance level–horizontal line on graph) with a influence statistic of 0.44 indicating the model is consistent with the magnitude and influence of the residuals.
We thank Terrie Taylor of the Blantyre Malaria Project, Blantyre, and Mark Drayson of the University of Birmingham for helpful advice on clinical and immunological aspects of the study. We are grateful to Nicholas Feasey, Robert Heyderman and Ian MacLennan for critical review of the manuscript. We thank the parents, guardians and children who participated in this study, and the staff at Queen Elizabeth Central Hospital for their assistance.
- 1. Berkley JA, Lowe BS, Mwangi I, Williams T, Bauni E, Mwarumba S, et al. Bacteremia among children admitted to a rural hospital in Kenya. N Engl J Med 2005; 352: 39–47. pmid:15635111
- 2. Reddy EA, Shaw AV, Crump JA. Community-acquired bloodstream infections in Africa: a systematic review and meta-analysis. Lancet Infect Dis 2010; 10: 417–432. pmid:20510282
- 3. Sigaúque B, Roca A, Mandomando I, Morais L, Quintó L, Sacarlal J, et al. Community-acquired bacteremia among children admitted to a rural hospital in Mozambique. Pediatr Infect Dis J 2009; 28: 108–113. pmid:19131902
- 4. Walsh AL, Phiri AJ, Graham SM, Molyneux EM, Molyneux ME. Bacteremia in febrile Malawian children: clinical and microbiologic features. Pediatr Infect Dis J 2000; 19: 312–318. pmid:10783021
- 5. Enwere G, Cheung YB, Zaman SM, Akano A, Oluwalana C, Brown O, et al. Epidemiology and clinical features of pneumonia according to radiographic findings in Gambian children. 2007; 12: 1377–1385. pmid:18045264
- 6. MacLennan CA, Martin LB, Micoli F. Vaccines against invasive Salmonella disease: Current status and future directions. Hum Vaccin Immunother 2014; 10: 1478–1493. pmid:24804797
- 7. MacLennan CA, Levine MM. Invasive nontyphoidal Salmonella disease in Africa: current status. Expert Rev Anti Infect Ther 2013; 11: 443–446. pmid:23627848
- 8. Uche IV, MacLennan CA, Saul A. A systematic review of the incidence, risk factors and case fatality rates of invasive nontyphoidal Salmonella (iNTS) disease in Africa (1966 to 2014). PLoS Negl Trop Dis 2017; 11: e0005118. pmid:28056035
- 9. Brent AJ, Oundo JO, Mwangi I, Ochola L, Lowe B, Berkley JA. Salmonella bacteremia in Kenyan children. The Pediatric infectious disease journal 2006; 25: 230–236. pmid:16511385
- 10. Mandomando I, Macete E, Sigauque B, Morais L, Qunitó L, Sacarlal J, et al. Invasive non-typhoidal Salmonella in Mozambican children. Trop Med Int Health 2009; 14: 1467–1474. pmid:19793081
- 11. Vandenberg O, Nyarukweba DZ, Ndeba PM,Hendriksen RS, Barzilay EJ, Schirvel C, et al. Microbiologic and clinical features of Salmonella species isolated from bacteremic children in eastern Democratic Republic of Congo. Pediatr Infect Dis J 2010; 29: 504–510. pmid:20104200
- 12. Tabu C, Breiman RF, Ochieng B, Aura B, Cosmas L, Audi A, et al. Differing burden and epidemiology of non-Typhi Salmonella bacteremia in rural and urban Kenya, 2006–2009. PLoS One 2012; 7: e31237. pmid:22363591
- 13. Mtove G, Amos B, von Seidlein L,Hendriksen I, Mwambuli A, Kimera J, et al. Invasive salmonellosis among children admitted to a rural Tanzanian hospital and a comparison with previous studies. PLoS One 2010; 16: e9244.
- 14. Feasey NA, Dougan G, Kingsley RA, Heyderman RS, Gordon MA. Invasive non-typhoidal salmonella disease: an emerging and neglected tropical disease in Africa. Lancet 2012; 379: 2489–2499. pmid:22587967
- 15. Molyneux EM, Mankhambo LA, Phiri A, Graham SM, Forsyth H, Phiri A, et al. The outcome of non-typhoidal salmonella meningitis in Malawian children, 1997–2006. Ann Trop Paediatr 2009; 29: 13–22. pmid:19222929
- 16. Hohmann EL. Nontyphoidal salmonellosis. Clin Infect Dis 2001; 32:263–9. pmid:11170916
- 17. Scott JA, Berkley JA, Mwangi I, Ochola L, Uyoga S, Macharia A, et al. Relation between falciparum malaria and bacteraemia in Kenyan children: a population-based, case-control study and a longitudinal study. Lancet 2011; 378: 1316–1323. pmid:21903251
- 18. Mackenzie G, Ceesay SJ, Hill PC,Walther M, Bojang KA, Satoguina J, et al. A decline in the incidence of invasive non-typhoidal Salmonella infection in The Gambia temporally associated with a decline in malaria infection. PLoS One 2010; 5: e10568. pmid:20485496
- 19. Mtove G, Amos B, Nadjm B, Hendriksen IC, Dondorp AM, Mwambuli A, et al. Decreasing incidence of severe malaria and community-acquired bacteraemia among hospitalized children in Muheza, north-eastern Tanzania, 2006–2010. Malar J 2011; 10: 320. pmid:22029477
- 20. Tapia MD, Tennant SM, Bornstein K, Onwuchekwa U, Tamboura B, Maiga A, et al. Invasive Nontyphoidal Salmonella Infections Among Children in Mali, 2002-2014: Microbiological and Epidemiologic Features Guide Vaccine Development. Clin Infect Dis. 2015; 61 Suppl 4:S332–8.
- 21. Kwambana-Adams B, Darboe S, Nabwera H, Foster-Nyarko E, Ikumapayi UN, Secka O, et al. Salmonella Infections in The Gambia, 2005–2015. Clin Infect Dis. 2015; 61 Suppl 4:S354–62.
- 22. Kalonji LM, Post A, Phoba MF, Falay D, Ngbonda D, Muyembe JJ, et al. Invasive Salmonella Infections at Multiple Surveillance Sites in the Democratic Republic of the Congo, 2011–2014. Clin Infect Dis. 2015; 61 Suppl 4: S346–53.
- 23. Oneko M, Kariuki S, Muturi-Kioi V, Otieno K, Otieno VO, Williamson JM, et al. Emergence of Community-Acquired, Multidrug-Resistant Invasive Nontyphoidal Salmonella Disease in Rural Western Kenya, 2009–2013. Clin Infect Dis. 2015; 61 Suppl 4: S310–6.
- 24. Verani JR, Toroitich S, Auko J, Kiplang'at S, Cosmas L, Audi A, et al. Burden of Invasive Nontyphoidal Salmonella Disease in a Rural and Urban Site in Kenya, 2009-2014. Clin Infect Dis. 2015; 61 Suppl 4: S302–9.
- 25. Mandomando I, Bassat Q, Sigaúque B, Massora S, Quintó L, Ácacio S, et al. Invasive Salmonella Infections Among Children From Rural Mozambique, 2001–2014. Clin Infect Dis. 2015; 61 Suppl 4: S339–45.
- 26. Keddy KH, Musekiwa A, Sooka A, Karstaedt A, Nana T, Seetharam S, et al. Clinical and microbiological features of invasive nontyphoidal Salmonella associated with HIV-infected patients, Gauteng Province, South Africa. Medicine 2017; 96: e6448. pmid:28353576
- 27. Agnandji ST, Lell B, Soulanoudjingar SS, Fernandes JF, Abossolo BP, Conzelmann C, et al. First results of phase 3 trial of RTS,S/AS01 malaria vaccine in African children. N Engl J Med 2011; 365: 1863–1875. pmid:22007715
- 28. Marks F, von Kalckreuth V, Aaby P, Adu-Sarkodie Y, El Tayeb MA, Ali M, et al. Incidence of invasive salmonella disease in sub-Saharan Africa: a multicentre population-based surveillance study. Lancet Glob Health 2017; 5:e310–e323. pmid:28193398
- 29. Mabey DC, Brown A, Greenwood BM. Plasmodium falciparum malaria and Salmonella infections in Gambian children. J Infect Dis 1987; 155: 1319–1321. pmid:3553352
- 30. Bronzan RN, Taylor TE, Mwenechanya J, Tembo M, Kayira K, Bwanaisa L, et al. Bacteremia in Malawian children with severe malaria: prevalence, etiology, HIV coinfection, and outcome. J Infect Dis 2007; 195: 895–904. pmid:17299721
- 31. Feasey NA, Everett D, Faragher EB, Roca-Feltrer A, Kang'ombe A, Denis B, et al. Modelling the Contributions of Malaria, HIV, Malnutrition and Rainfall to the Decline in Paediatric Invasive Non-typhoidal Salmonella Disease in Malawi. PLoS Negl Trop Dis. 2015; 9:e0003979. pmid:26230258
- 32. Bachou H, Tylleskar T, Kaddu-Mulindwa DH, Tumwine JK. Bacteraemia among severely malnourished children infected and uninfected with the human immunodeficiency virus-1 in Kampala, Uganda. BMC Infect Dis 2006; 6: 160. pmid:17090299
- 33. Calis JC, Phiri KS, Faragher EB, Brabin BJ, Bates I, Cuevas LE, et al. Severe anemia in Malawian children. N Engl J Med 2008; 358: 888–899. pmid:18305266
- 34. Gordon MA, Banda HT, Gondwe M, Gordon SB, Boeree MJ, Walsh AL, et al. Non-typhoidal salmonella bacteraemia among HIV-infected Malawian adults: high mortality and frequent recrudescence. AIDS 2002; 16: 1633–1641. pmid:12172085
- 35. Gordon MA, Graham SM, Walsh AL,Wilson L, Phiri A, Molyneux E, et al. Epidemics of invasive Salmonella enterica serovar enteritidis and S. enterica Serovar typhimurium infection associated with multidrug resistance among adults and children in Malawi. Clin Infect Dis 2008; 46: 963–969. pmid:18444810
- 36. Fleming S, Thompson M, Stevens R, Heneghan C, Plüddemann A, Maconochie I, et al. Normal ranges of heart rate and respiratory rate in children from birth to 18 years of age: a systematic review of observational studies. Lancet 2011; 377: 1011–1018. pmid:21411136
- 37. Graham SM, Walsh AL, Molyneux EM, Phiri AJ, Molyneux ME. Clinical presentation of non-typhoidal Salmonella bacteraemia in Malawian children. Trans R Soc Trop Med Hyg 2000; 94: 310–314. pmid:10975008
- 38. Graham SM, English M. Non-typhoidal salmonellae: a management challenge for children with community-acquired invasive disease in tropical African countries. Lancet 2009; 373: 267–269. pmid:19150705
- 39. Feasey NA, Masesa C, Jassi C, Faragher EB, Mallewa J, Mallewa M, et al. Three epidemics of multidrug resistant Salmonella bloodstream infection in Blantyre, Malawi. Clin Infect Dis 2015; 61 Suppl 4: S363–371.
- 40. Okoro CK, Kingsley RA, Connor TR, Harris SR, Parry CM, et al. Epidemic intra-continental spread of human invasive Salmonella Typhimurium pathovariants in sub-Saharan Africa. Nature Genetics 2012; 44: 1215–1221. pmid:23023330
- 41. Feasey NA, Hadfield J, Keddy KH, Dallman TJ, Jacobs J, Deng X, et al. Distinct Salmonella Enteritidis lineages associated with enterocolitis in high-income settings and invasive disease in low-income settings. Nat Genet 2016; 48: 1211–1217. pmid:27548315
- 42. Cunnington AJ, de Souza JB, Walther M, Riley EM. Malaria impairs resistance to Salmonella through heme- and heme oxygenase-dependent dysfunctional granulocyte mobilization. Nat Med 2012; 18: 120–127.
- 43. Cunnington AJ, Njie M, Correa S, Takem EN, Riley EM, Walther M. Prolonged neutrophil dysfunction after Plasmodium falciparum malaria is related to hemolysis and heme oxygenase-1 induction. J Immunol 2012; 189: 5336–5346. pmid:23100518
- 44. UNICEF. The state of the world's children 2007. Available at: http://www.unicef.org/publications/files/The_State_of_the_Worlds__Children__2007_e.pdf. Accessed 8 October 2017.
- 45. MacLennan CA, Gondwe EN, Msefula CL, Kingsley RA, Thomson NR, White SA, et al. The neglected role of antibody in protection against bacteremia caused by nontyphoidal strains of Salmonella in African children. J Clin Invest 2008; 118: 1553–1562. pmid:18357343
- 46. Gondwe EN, Molyneux ME, Goodall M, Graham SM, Mastroeni P, Drayson MT, et al. Importance of antibody and complement for oxidative burst and killing of invasive nontyphoidal Salmonella by blood cells in Africans. Proc Natl Acad Sci U S A 2010; 107: 3070–3075. pmid:20133627
- 47. Nyirenda TS, Gilchrist JJ, Feasey NA, Glennie SJ, Bar-Zeev N, Gordon MA, et al. Sequential acquisition of T-cells and antibodies to nontyphoidal Salmonella in Malawian Children. J Infect Dis 2014; 210: 56–64. pmid:24443544
- 48. Raffatellu M, Santos RL, Verhoeven DE, George MD, Wilson RP, Winter SE, et al. Simian immunodeficiency virus-induced mucosal interleukin-17 deficiency promotes Salmonella dissemination from the gut. Nat Med 2008; 14: 421–428. pmid:18376406
- 49. Gordon MA, Gordon SB, Musaya L, Zijlstra EE, Molyneux ME, Read RC. Primary macrophages from HIV-infected adults show dysregulated cytokine responses to Salmonella, but normal internalization and killing. AIDS 2007; 21: 2399–2408. pmid:18025876
- 50. MacLennan CA, Gilchrist JJ, Gordon MA, Cunningham AF, Cobbold M, Goodall M, et al. Dysregulated humoral immunity to nontyphoidal Salmonella in HIV-infected African adults. Science 2010; 328: 508–512. pmid:20413503
- 51. Goh YS, Necchi F, O’Shaughnessy CM, Micoli F, Gavini M, Young S, et al. Bactericidal Immunity to Salmonella in Africans and Mechanisms Causing its failure in HIV Infection. 2016 PLoS Negl Trop Dis. 2016; 10: e30004604.
- 52. Wiktor SZ, Sassan-Morokro M, Grant AD, Abouya L, Karon JM, Maurice C, et al. Efficacy of trimethoprim-sulphamethoxazole prophylaxis to decrease morbidity and mortality in HIV-1-infected patients with tuberculosis in Abidjan, Cote d'Ivoire: a randomised controlled trial. Lancet 1999; 353: 1469–1475. pmid:10232312
- 53. Zachariah R, Harries AD, Luo C, Bachman G, Graham SM. Scaling-up co-trimoxazole prophylaxis in HIV-exposed and HIV-infected children in high HIV-prevalence countries. Lancet Infect Dis 2007; 7: 686–693. pmid:17897611