A Perspective on Invasive Salmonella Disease in Africa Free

Abstract

Salmonellaenterica is a leading cause of community-acquired bloodstream infection in Africa. The contribution of typhoidal and nontyphoidal Salmonella serovars to invasive disease varies considerably in place and time, even within the same country. Nonetheless, many African countries are now thought to experience typhoid fever incidence >100 per 100 000 per year with approximately 1% of patients dying. Invasive nontyphoidal Salmonella (iNTS) disease was estimated to cause 3.4 million illnesses and 681 316 deaths in 2010, with the most disease in Africa. Antimicrobial drug resistance is a growing problem in S. enterica that threatens to further compromise patient outcomes. Reservoirs for nontyphoidal Salmonella and the predominant routes of transmission for typhoidal and nontyphoidal Salmonella are not well understood in Africa, hampering the design of evidence-based, non-vaccine- and vaccine-based prevention measures. It is difficult to distinguish clinically invasive Salmonella disease from febrile illnesses caused by other pathogens. Blood cultures are the mainstay of laboratory diagnosis, but lack sensitivity due to the low magnitude of bacteremia, do not produce results at point of care, and are not widely available in Africa. Serologic approaches to diagnosis remain inaccurate, and nucleic acid amplification tests are also compromised by low concentrations of bacteria. High-throughput whole-genome sequencing, together with a range of novel analytic pipelines, has provided new insights into the complex pattern of epidemiology, pathogenesis, and host adaptation. Concerted efforts are therefore needed to apply these new tools in the context of high-quality field surveillance to improve diagnosis, patient management, control, and prevention of invasive Salmonella infections in Africa.

Salmonella enterica is a leading cause of community-acquired bloodstream infections in both Africa and Asia [1, 2]. Salmonella enterica serovars Typhi, Paratyphi A, Paratyphi B, and Paratyphi C may be referred to collectively as typhoidal Salmonella, whereas other serovars are grouped as nontyphoidal Salmonella (NTS). Typhoidal Salmonella strains are human host–restricted organisms that cause typhoid fever and paratyphoid fever, together referred to as enteric fever. NTS strains may be host generalists, infecting or colonizing a broad range of vertebrate animals, or may be adapted or restricted to particular nonhuman animal species [3].

HISTORIC BACKGROUND

Since the early use of blood cultures in Africa, invasive S. enterica infections appear to have been common [4, 5]. However, the limited distribution and use of blood culture services in many African countries has prevented the development of a clear picture of patterns of disease [6], a situation that has only begun to change relatively recently [1]. Across the continent, lack of sustained progress with access to improved water and sanitation facilities [7], combined with rapid and unplanned urbanization, create the conditions that favor typhoid fever transmission [8]. Recent estimates suggest that typhoid fever incidence and outbreaks may have increased in Africa over the past decade [9]. Large populations of immunocompromised adults and children appear to have favored the emergence of invasive nontyphoidal Salmonella (iNTS) disease [3]. Indeed, NTS is a leading cause of invasive bacterial disease in many countries in sub-Saharan Africa [1]. Host factors appear to be important in risk for iNTS disease. An association between malaria and invasive Salmonella infections was suggested in 1929 [10] and confirmed in 1987 [11]. Today, malaria and malnutrition in infants and children, and human immunodeficiency virus (HIV) infection in adults, are consistently identified as important host risk factors for iNTS that are all disproportionately more common in African countries compared to other regions [3]. Of great concern, antimicrobial resistance among iNTS strains in Africa is increasing both in prevalence and number of antimicrobial classes compromised, including extended-spectrum cephalosporins [12], threatening existing empiric management regimens for severe febrile illness and sepsis.

BURDEN OF DISEASE

In 2000 it was estimated that Salmonella Typhi incidence in Africa was 50 illnesses per 100 000 persons per year with a case fatality ratio of 1% [13]. However, this estimate was based on very limited source data from the control arms of 2 vaccine trials in Egypt and 1 in South Africa conducted a decade or more earlier. Since then, a number of new typhoid incidence studies have been published from other African countries. These data have contributed to a 2010 estimate of typhoid burden that suggests that most of sub-Saharan Africa now experiences typhoid fever incidence >100 per 100 000 persons per year, with 33 490 deaths, 26.0% of global typhoid deaths, occurring in Africa [9]. Multidrug-resistant (MDR) Salmonella Typhi strains of the H58 haplotype that probably originated in Asia are now spreading on the continent [14], which is likely to herald greater risk for death when ineffective antimicrobials are used. The global burden of iNTS disease was estimated for the first time for the year 2010 [15]. Of the 3.4 million cases and 681 316 deaths in 2010 due to iNTS disease worldwide, the majority of these were thought to occur in Africa. Here, infants and young children living in areas of high malaria transmission intensity or with severe acute malnutrition and young adults with HIV are at particular risk of iNTS [3]. However, despite declines in malaria [16], iNTS disease persists as a problem. Perhaps due to comorbidities in many patients, the case fatality ratio of iNTS disease appears to be an order of magnitude higher than for typhoid fever, placing it among some of the most deadly infectious diseases on the continent.

Future burden of disease estimates will be refined as more population-based studies report the incidence of invasive Salmonella infections. Some such studies are published in this supplement, and others are expected soon, including a multicountry study of typhoid fever incidence in Africa [17]. However, in addition to refining estimates of incidence, a better understanding of disability and deaths, especially outside of the hospital, is needed to inform investment decisions on prevention and control [18, 19].

EPIDEMIOLOGY

Although S. enterica appears to be a leading cause of community-acquired bloodstream infections in most places in Africa where the epidemiology of bacteremia has been studied [1], the relative contributions of typhoidal and nontyphoidal serovars vary in both place [20] and time, even within countries [21, 22]. African countries have lagged behind those in other regions in access to improved water and sanitation facilities [7]. Humans are the source of Salmonella Typhi worldwide, and transmission is usually by the waterborne and foodborne routes. However, there have been few epidemiologic studies to establish locally important risk factors for typhoid fever transmission in either rural or urban settings in Africa. The predominant NTS strain causing invasive disease in Africa, Salmonella Typhimurium ST313, appears to show genomic features of differential host adaptation and convergent evolution with Salmonella Typhi. This raises the possibility that humans may be an important source of infection for this ST313 strain [23]. However, the predominant routes of transmission of NTS in Africa remain to be defined. Salmonella Paratyphi A, B, and C have so far been uncommon causes of bacteremia in Africa [1], but this needs to be monitored closely. In-depth field epidemiologic research is needed to refine both vaccine and nonvaccine approaches to the prevention and control of invasive Salmonella disease in Africa.

CHALLENGES TO ESTABLISHED DOGMA

Over many years, assumptions regarding the epidemiology, pathogenesis, and clinical manifestations of typhoid fever and iNTS disease in sub-Saharan Africa have been extrapolated from experience elsewhere in the world. In relation to typhoid fever, it has been widely assumed that the incidence and mortality is low; that disease principally occurs in older children and adults; that multidrug resistance, although a pressure on resources, has little effect on epidemiology; and that Salmonella Typhi within households is due to a common source rather than multiple sources. Likewise, there is the popular view that chronic carriers are the major reservoir of infection in the population. In relation to NTS, it has been assumed that clinical illness due to NTS is most commonly self-limited diarrhea; that animals are the reservoir for the common Salmonella Typhimurium and Salmonella Enteritidis serovars; that transmission of NTS is mainly from animal contaminated foods and water; and that NTS strains are equal in their disease profile, in particular their invasiveness. However, in recent years, new evidence has emerged that has challenged many of these concepts. What has made the difference has been the establishment of long-term laboratory-based surveillance at several African sites, careful clinical description of patients with iNTS disease, and the capacity for large-scale whole-genome sequencing (WGS). This supplement will present some of the key aspects of these novel data.

INSIGHTS FROM BACTERIAL GENETICS

The advent of high-throughput WGS together with a range of novel analytic pipelines has shown how the complex pattern of epidemiology and host adaptation among different Salmonella serovars is reflected in the genome [24]. The Salmonella Typhi genome was first reported in 2001 [25], and since then multiple Salmonella serovars and several thousand Salmonella strains have been sequenced. In relation to Salmonella Typhi, we have learned that, in keeping with being limited to a human host, this serovar has a restricted genetic repertoire with considerable gene inactivation. Intriguingly, genome degradation in the form of pseudogenes and deletions has also been seen in Salmonella Typhimurium, one of the most common Salmonella serovars associated with iNTS disease in Africa [23]. Indeed, these invasive strains largely come from a single multilocus sequence type (a strain typing method based on the profile of housekeeping genes), Salmonella Typhimurium ST313, in which some of the affected genes are related to pathogenesis and others to anaerobic and enteric metabolism, but many are of unknown function. Convergence between Salmonella Typhimurium ST313 and Salmonella Typhi raises the possibility that ST313 has adapted to occupy a unique niche in Africa with a different host range. By providing much greater resolution of phylogeny than MLST and facilitating the reconstruction of evolutionary history using Bayesian algorithms, WGS has also revealed putative pathways and mechanisms of the spread of Salmonella Typhi and Salmonella Typhimurium across the African continent [14, 26, 27], and the location of transposons encoding multidrug resistance in plasmids that in some strains have become fixed within the chromosome. Further molecular analyses of emerging MDR strains of Salmonella Typhi, invasive and intestinal strains of Salmonella Typhimurium, and other common invasive disease-causing serovars in Africa such as Salmonella Enteritidis and Salmonella Dublin are currently under way. With the appropriate epidemiologic context, it is hoped that these studies will help unravel the mysteries of disease reservoirs and modes of transmission, and help to predict the emergence of antimicrobial resistance.

UNDERSTANDING PATHOGENESIS

In Africa, NTS strains appear to be different from those that cause diarrheal disease in industrialized countries, in that they more often cause invasive disease with bacteremia. Immunocompromised hosts, including young children who are malnourished, HIV infected, or with malaria-associated anemia or adults with advanced HIV infection, are at particular risk of developing iNTS disease [3]. Little is known about the pathogenesis of NTS in these populations. It is thought that intestinal mucosal and subsequent bloodstream invasion occur as a result of defects in immune surveillance, at the level of both cellular and antibody-mediated immunity [28–30]. We know that NTS invades the bone marrow, which may act as a sanctuary site. However, mechanisms of systemic dissemination, reservoirs of carriage, control of relapse, and mechanisms for onward transmission have not been well defined. Further experiments based on the recent studies demonstrating severe disease in animal infections with Salmonella Typhimurium ST313 will help address some of these complexities [31]. In contrast, the mechanisms underlying typhoidal systemic disease, in particular the ability of Salmonella Typhi to avoid inflammation and translocate to sanctuary sites, are better understood. The identification of a typhoid toxin [32], an AB-like toxin that binds to specific human host cell glycoproteins and causes typhoid fever in animal models, may provide a functional explanation for both the host restriction and the clinical manifestations of the disease. Mechanisms of gallbladder colonization and persistence have increasingly been better described in recent years, revealing how gallstones act as a substrate for Salmonella biofilm formation [33]. However, why immunocompromise due to HIV and malnutrition, for example, do not appear to be risk factors for invasive Salmonella Typhi disease in Africa remains uncertain [1, 20]. In addition to the bacterial genetic studies discussed previously, further investigation of this apparent paradox and the interaction of Salmonella with the human gut immune system will help us better understand the differences in the biology of Salmonella Typhi vs NTS serovars.

CHALLENGES IN DIAGNOSIS

Invasive Salmonella infections present most often as a febrile illness that is difficult to distinguish clinically from other febrile diseases including malaria [34, 35]. Blood culture remains the standard for diagnosis, but, like other culture-based methods, is not widely available in many African countries [36] and does not produce results in time to inform the initial management of individual patients. Bone marrow culture is more sensitive for the diagnosis of typhoid fever but is less practical than blood culture. The median magnitude of bacteremia in Salmonella bloodstream infection is <1 colony forming unit/ mL [37, 38], contributing to the relatively low sensitivity of blood culture and also hampering molecular approaches to diagnosis [39]. The low magnitude of bacteremia can be compensated for, at least in part, by inoculating larger blood volumes for culture or extracting larger volumes for nucleic acid amplification. Serologic approaches to diagnosis, ranging from the Widal agglutination test [40] to enzyme-linked immunosorbent assays and point-of-care diagnostics [41], have shortcomings of either sensitivity, specificity, or both that preclude their use for individual patient management [42, 43].

Progress is needed to make culture-based diagnostics more widely available at district hospitals in Africa [6], while progressing efforts to develop non-culture-based diagnostic tests that are useful at both the hospital and health center level, and in more remote environments. Such efforts are needed not only to improve individual patient diagnosis and outcomes, but also to develop strategies to assess population exposure to refine burden of disease estimates and to better target prevention measures, including vaccines.

MANAGEMENT CHALLENGES INCLUDING THE THREAT OF ANTIMICROBIAL RESISTANCE

Historically, iNTS and typhoid disease deaths have been largely captured in hospital settings. A better understanding of the size of the catchment population from which hospitalized cases are drawn, what proportion of all cases present at the hospital, and improved capture of vital events in the community are essential if we are to derive more accurate estimates of disease burden. Clinical recognition of invasive Salmonella disease remains a challenge where both typhoid and NTS occur [3]. Given the clear differences in their genetics and pathogenesis, it is not safe to assume that all invasive Salmonella, particularly in Africa, can be managed in the same way. The lack of robust diagnostics and the poor availability and sensitivity of blood cultures lead to frequent empiric treatment, sometimes with inappropriate antimicrobials guided by national or international guidelines [44, 45]. As discussed, older serology-based tests have proven too nonspecific to reliably guide either therapy or epidemiology. Novel targets have been evaluated but need to be incorporated into more user-friendly applications and field tested in environments of diverse transmission intensity.

The emergence of antimicrobial resistance is alarming in several different ways. Multidrug resistance to antimicrobial agents commonly available in community settings is already highly prevalent. This is likely to result in hospitals treating larger numbers of more severely ill patients who have failed to respond to inadequate initial therapy. Antimicrobial resistance may also be associated with increased case fatality ratios and longer transmission times due to inadequate clearance driving higher incidence [46]. Reports of NTS and Salmonella Typhi in Africa resistant to extended-spectrum cephalosporins and fluoroquinolones, the identification of transmissible mobile genetic elements containing the resistance genes, and the fixation of some antimicrobial resistance genes in the chromosome raise the possibility of rapid changes in the burden of disease in the future [14, 22, 23]. As discussed, mechanisms to monitor the emergence of antimicrobial resistance need to be better defined and are essential if we are going to be able to move swiftly enough to control disease, particularly in the context of the large outbreaks that have been experienced in recent years.

FUTURE DIRECTIONS

Concerted efforts are needed if the scientific, public health, and policy making communities are going to overcome the current high burden of iNTS disease and typhoid fever in Africa. These efforts will need to start with improved diagnostics that will then underpin the high-quality field and molecular epidemiology required to inform novel approaches to disease modeling. Such models may help to fill challenging gaps in our knowledge of disease source and transmission, and have the potential to predict the impact of nonvaccine and vaccine interventions. Surveillance to detect further development of antimicrobial resistance and trials to establish optimal case management are essential. Improving patient outcomes will require clinical studies to inform patient management algorithms. Although we are still largely ignorant of the burden of disease, potential risk factors, routes of transmission, epidemiological trends, and reservoirs of infection, each of these elements is addressed in this supplement. The momentum for the supplement arose as a result of a consensus meeting of experts on invasive Salmonella in Africa held in Blantyre, Malawi, at the end of 2014 [47]. Sponsored by the Wellcome Trust and the Bill & Melinda Gates Foundation, new data were presented from 17 African countries, resulting in this supplement, which will help bridge critical knowledge gaps and spur new collaborations between multidisciplinary teams across continents. It is also hoped that the supplement will promote advocacy at the policy-maker and funder levels to improve the visibility of these neglected diseases.

Notes

Financial support. Publication of this paper was made possible with the support of the Bill & Melinda Gates Foundation (grant number OPP1125993). J. A. C. is supported by the joint US National Institutes of Health–National Science Foundation Ecology and Evolution of Infectious Disease program (grant number R01 TW009237) and the UK Biotechnology and Biological Sciences Research Council (BBSRC; grant number BB/J010367/1), and by the UK BBSRC Zoonoses in Emerging Livestock Systems program (award numbers BB/L017679, BB/L018926, and BB/L018845). R. S. H. was supported by a Strategic Award from the Wellcome Trust for the Malawi-Liverpool-Wellcome Trust Clinical Research Programme.

Supplement sponsorship. This article appeared as part of the supplement “Invasive Salmonella Disease in Africa,” sponsored by the University of Otago.

Potential conflict of interest. Both authors: No reported conflicts.

Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References