Update on the burden of Campylobacter in developing countries

Abstract

Purpose of review

Recent work has added to the understanding of the burden of Campylobacter jejuni, C. coli, and non-jejuni/coli Campylobacter strains in children living in the developing world.

Recent findings

New diagnostic modalities and carefully-designed field studies are demonstrating that the burden of Campylobacter diarrhea in children in the developing world has been greatly underestimated. Furthermore, there is emerging recognition of an association between Campylobacter infection and malnutrition. Important progress has been made towards a Campylobacter jejuni vaccine. Finally, evidence of antibiotic resistance continues to be an important issue that is accentuated by the realization that the burden of disease is greater than previously recognized.

Summary

Additional research is needed to refine our understanding of the epidemiology of Campylobacter infections in developing countries, in particular to improve estimates of the burden of Campylobacter diarrhea in endemic settings, to determine the impact of recurrent Campylobacter infections on child development, and to describe the prevalence and clinical significance of non-jejuni/coli Campylobacter infections. Progressive antibiotic resistance of isolates argues for augmented and expanded control measures of antibiotics in livestock. Continued work in vaccine development is warranted as is the extension of data available on the serotypes related to burden in different areas of the world and the relationship of serotypes to disease severity.

INTRODUCTION

Campylobacter is considered to be the most common cause of bacterial gastroenteritis worldwide and was associated with 7.5 million disability-adjusted life years in the 2010 Global Burden of Disease Study, more than Shigella (7.1 million) and enterotoxigenic E. coli (6.9 million)[1]. The epidemiology of Campylobacter infection in developed countries is notably different than in the developing world. In the former, Campylobacter infections are sporadic, the prevalence of asymptomatic infection is low, and marked seasonality is seen [2]. With the exception of common-source outbreaks, often associated with contaminated dairy products [3], infections are typically linked directly to contaminated food, including poultry [2]. In contrast, in developed countries where Campylobacter is often endemic, asymptomatic infections are common and seasonality is less prominent or even absent [4-6]. Campylobacter diarrhea is seen in children but rarely in adults [7]. Due to the ubiquitous nature of the pathogen, risk factors are more diffusely associated with exposure to the environment, including contaminated drinking water [8].

While the burden of diarrhea in developed countries has been established by longitudinal surveillance [9,10], accurate estimates of the burden of diarrhea as well as long-term sequelae of Campylobacter infection in the developing world remain elusive, in part due to insensitive and inconsistently applied isolation methods; this forms a poor lens through which to study the burden of this common enteropathogen. Studies with improved diagnostics remain limited in number and epidemiologic context, but it is increasingly clear that current estimates significantly underestimate disease burden. Here, we will review developments in diagnostics, studies of both the acute and long-term disease burden, non-jejuni/coli infections, and new insights on Campylobacter immunity and vaccine development. We end by reviewing new data on antibiotic-resistant Campylobacter infections.

UPDATE ON CAMPYLOBACTER DIAGNOSTICS

Campylobacter species are difficult to isolate in the presence of normal stool flora, thus selective techniques are commonly employed. This most commonly involves the use of antibiotic-containing selective media, which has been the primary diagnostic modality for defining the epidemiology of Campylobacter jejuni, the species most commonly implicated in human disease. C. jejuni isolation is also facilitated by modified thermal and atmospheric conditions [11]. However, both the culture media and conditions can inhibit isolation of other Campylobacter species, and alternative strategies including membrane filtration culture and molecular diagnostics have led to increased detection of a broad range of Campylobacter species [12-14]. In South Africa, using filtration culture, Campylobacter jejuni isolates represent only about 40% of all Campylobacter isolates from children with diarrhea [15]. However, this approach has not been widely adopted.

Over the last decade, culture-independent tests have been developed, including enzyme immunoassay (EIA)-based stool antigen tests and nucleic acid-based tests. Early studies suggested that antigen tests had excellent sensitivity and specificity in comparison to culture in children and adults in the developed world [16]. However with routine use in clinical microbiology laboratories, several recent reports have described a high false-positivity rate [17,18]. Ghosh and colleagues performed culture, EIA and PCR for C. jejuni/coli on stool specimens from more than 500 children with acute diarrhea presenting to a tertiary care hospital in New Delhi and found culture to have a sensitivity of 37% compared to a gold standard of either culture positive or both EIA and PCR positive [19]. There is now good evidence that these antigen tests detect non-jejuni/coli Campylobacter species, including C. upsaliensis, C. concisus, C. troglodytis, C. hyointestinalis subsp. hyointestinalis, C. helveticus and C. lanienae [20]. PCR-based tests for detection of Campylobacter species directly from stool as well as genus-level assays have begun to be used more widely on clinical specimens, including amplification of 16S ribosomal RNA targets with subsequent sequencing for species identification [14,20,21].

BURDEN OF CAMPYLOBACTER GASTROENTERITIS

Population-based surveillance for Campylobacter in the developing world is limited by the frequent asymptomatic detection of Campylobacter in stools, which complicates attribution when Campylobacter is identified in studies of symptomatic gastroenteritis. Consequently, controlled etiologic studies of diarrhea in these settings have been employed to improve burden estimates. The recently published Global Enteric Multicenter Study (GEMS) of moderate-to-severe diarrhea in children in Asia and Africa was designed to comprehensively identify the pathogen-specific etiology and burden of disease associated with each pathogen [22]. Seven sites in Africa and Asia enrolled cases of moderate-to-severe acute diarrhea and asymptomatic controls matched for age, sex, and village, and in addition to a broad range of other pathogens, Campylobacter was detected by culture using selective blood agar plates incubated at 42°C [23]. C. jejuni was found to be associated with a significant burden of diarrhea in Mirzapur, Bangladesh (fourth highest in the first year of life behind rotavirus, Shigella and Aeromonas), Karachi, Pakistan (fifth highest in the first year of life behind rotavirus, Aeromonas, Shigella and heat-stable toxin-producing enterotoxigenic E. coli and second highest in children ages 2-5 behind Aeromonas) and Kolkata, India (third highest in children ages 2-5 behind rotavirus and Shigella) but not in four African sites (in the Gambia, Mali, Mozambique, and Kenya). The high burden of Campylobacter diarrhea in the oldest age group at some GEMS sites is surprising in light of Calva et al.'s classical evidence from a Mexican birth cohort that the ratio of symptomatic to asymptomatic infections shifted towards more frequent asymptomatic infection by 6 months of age [24].

Studies with culture-independent diagnostic tests have also revealed a substantial burden of disease. The ongoing Study of Etiology, Risk Factors, and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health (MAL-ED), designed to assess the impact of malnutrition and recurrent early enteropathogen infections on long-term developmental outcomes in children across eight sites in South America, sub-Saharan Africa, and Asia, has used an EIA for Campylobacter detection, chosen for ease of standardization across laboratories [25]. Campylobacter has the highest associated burden of diarrhea of all pathogens in children in the first year of life from Loreto, Peru and Venda, South Africa [26]. Two recent analyses of archived specimens using PCR for C. jejuni/coli have also revealed a substantial burden of associated diarrhea. In testing of samples from a birth cohort of infants in Bangladesh, C. jejuni/coli detection was associated with diarrhea and was estimated to be a probable cause of 11.3% of episodes, second to enteroaggregative E. coli [4]. In a large number of samples from hospitalized children with and without diarrhea in Malawi, Campylobacter jejuni/coli was detected in 21% of diarrheal episodes and was strongly associated with diarrhea [27]. Interestingly, C. coli represented more than one-third of detections in diarrhea samples.

It is difficult to draw conclusions about the burden of Campylobacter infection in settings where the prevalence of asymptomatic infection is high enough to preclude a statistically significant association between detection and disease. It is possible that such settings have reached a force of infection at which the point prevalence of asymptomatic infection would be expected to be as high as or higher than that of symptomatic infection. Such a scenario has been suggested for norovirus, which similarly has a high burden of sporadic disease in the developed world but has been inconsistently associated with diarrhea in children in endemic settings [22,28]. It is also likely that the difference in Campylobacter burden estimates between these settings are explained by host factors, infection with less pathogenic Campylobacter species (in the case of detection with EIA), and variation in the pathogenicity of circulating serotypes.

Accurate estimation of the burden of acute diarrheal disease in such settings will likely require novel analytic approaches. These include 1) quantitative PCR-based testing, where the quantity of pathogen detected may help distinguish between high-level infection that is associated with diarrhea and low-level infection of unclear significance [29,30], 2) longitudinal studies with intensive surveillance, molecular typing and force of infection modeling, and 3) incorporation of serologic assays of immunity to better understand and classify asymptomatic infections [31]. Of note, such approaches have been critical for understanding the burden of disease associated with malaria infection in hyperendemic settings [32].

ENDURING CONSEQUENCES OF CAMPYLOBACTER INFECTION

Pathogens that have been most clearly associated with physiologic injury extending beyond the acute diarrheal illness include Cryptosporidium and Shigella in the developing world, and Campylobacter, Shigella and Salmonella in the developed world. Despite the diagnostic limitations, high quality evidence supporting a link with physiologic insults enduring beyond the acute illness is strongest and most consistent across epidemiologic contexts for Campylobacter infections.

Guillain-Barré syndrome (GBS)

C. jejuni is the pathogen most frequently associated with the development of GBS, specifically the acute motor axonal neuropathy (AMAN) variant [33]. The drive to eradicate poliomyelitis has improved surveillance for acute flaccid paralysis. Several high quality studies have recently shown that GBS is the leading cause of acute flaccid paralysis (AFP) in Bangladesh, with an estimated incidence rate of 3.25 cases/100,000 children under 15 years of age, and Campylobacter infection appears to precipitate the majority of cases in this setting [34,35]. Although much attention has been paid to the burden of Campylobacter infection and associated GBS in the developed world, it is clear that a substantial burden is present in the developing world as well. Application of molecular diagnostics to stools collected during AFP surveillance may help provide better resolution of this burden.

Inflammatory Bowel Disease and reactive arthritis

Although the associations between Campylobacter infection and functional gastrointestinal disorders [36] and to a lesser degree inflammatory bowel disease [37] and celiac disease [38] are described in the United States and Europe, evaluating such relationships is complicated in settings where enteropathogens are ubiquitous. No data from developing world cohorts were considered appropriate for inclusion in a recent review of the associations between Campylobacter and reactive arthritis [39].

Malnutrition

The relationship between Campylobacter and malnutrition appears to be bi-directional. The observation has frequently be made that Campylobacter infections are more common in malnourished children [5,40], and a recent microbiome study from a malnourished child in India found that the family Campylobacteraceae was 35 times more prevalent than in a well-nourished child [41]. Evidence from gnotobiotic mouse models suggests that shifts in diet and microbiota alter colonization resistance to Campylobacter jejuni [42]. These studies suggest that host factors are important and indeed that Campylobacter may be a biomarker of enteropathy. However, a single recent longitudinal study has shown that both asymptomatic and symptomatic Campylobacter jejuni/coli infections are marginally associated with the acquisition of linear growth over the nine months subsequent to the episode, with more severe disease associated with greater growth deficits, suggesting that infection is not merely a bystander in this context [6]. Validation of this finding in other longitudinal studies with high incidences of Campylobacter infection and diarrhea would support the investigation of interventional approaches to improve infant growth in developing world settings.

BURDEN OF NON-JEJUNI/COLI CAMPYLOBACTER INFECTIONS

Wider adoption of less-selective isolation techniques as well as culture-independent diagnostics has led to frequent reports of the clinical significance of non-jejuni/coli Campylobacter infections. An association with gastroenteritis has been shown for a broad range of species, including C. concisus [43], C. upsaliensis [44], C. fetus [45], C. ureolyticus [46], and C. hyointestinalis [47]. Lastovica and colleagues have shown that using non-selective culture, C. concisus and C. upsaliensis can be isolated with a frequency close to that of C. jejuni from diarrheal stools from children in South Africa, together representing 47% of all Campylobacter detections [48]. Using molecular tests, C. concisus was detected more frequently than C. jejuni in children with gastroenteritis in Belgium [49] and appears to be associated with milder but more prolonged diarrhea than C. jejuni when detected by filtration culture in Denmark [43] and has also been associated with inflammatory bowel disease [50]. There is a pressing need for a more comprehensive assessment of the incidence and pathogenicity of these species in developing countries.

Using an EIA, Campylobacter has been the most frequently detected pathogen in the MAL-ED study in children 0-2 years of age. We have shown that about one-third of EIA detections in the Bangladesh, Peru and Tanzania cohorts represented non-jejuni/coli species, the majority of which were identified as C. hyointestinalis subsp. lawsonii or C. troglodytis [20]. The former is of porcine origin and was the most frequently identified species in Peru and Tanzania. C. troglodytis was most frequently identified in Bangladesh and is similar to an uncultured species identified 16S rRNA-based sequencing of stool from infants in Bangladesh [51]. The clinical significance of infections with these species is not known. An association between non-jejuni/coli Campylobacter infections and malnutrition has not been described. Accurate estimates of the relative prevalence and importance of non-jejuni/coli Campylobacter species will require broader use of less selective diagnostic modalities.

PROSPECTS FOR CONTROL OF CAMPYLOBACTER JEJUNI INFECTION

Control of exposure to Campylobacter in the developing world will be challenging, as contamination of food and water is common. Specific interventions, such as corralling chickens, have not been demonstrated to have a significant public health impact [52]. Despite the challenge of developing a vaccine for C. jejuni, there has been some recent progress. Studies in the late 1980s showed homologous protection using C. jejuni 81-176 [53], however this strain contains GM2 and GM3 ganglioside mimics, implicated in the pathogenesis of Guillain-Barré syndrome. In order to avoid this safety concern, a strain lacking ganglioside mimics was described (CG 8421)[54] but recently demonstrated a lack of homologous protection in a human challenge model [55]. In parallel work, the development of a capsular conjugate vaccine, which in principal could both be more immunogenic and avoid the subsequent risk of Guillain-Barré syndrome [56,57], has been shown to confer protection in non-human primates.

Although classically C. jejuni was typed by Lior or Penner typing strategies, the demonstration that the capsule was the serodeterminant of the Penner scheme and that the capsule structure play an important role in facilitating colonization [58] has led to epidemiologic studies to define the prevalence of diverse capsular types in regions of the world which are important for guiding development of a multivalent vaccine [59]. Additional studies such as these are warranted, because despite the disappointing results in the recent study with strain CG8421, prior human challenge studies support homologous protection, as does the epidemiology of infections in children in endemic areas. The development of multiplex PCR-based capsular typing [60] should help close these knowledge gaps.

ANTIBIOTIC RESISTANCE

Increasingly, the progressive spread of antibiotic-resistant Campylobacter strains is a concern, in particular in Southeast Asia and South America. Resistance against fluoroquinolones has increased substantially over the last decade [61,62] and macrolide resistance has been increasingly described [63-65]. Although treatment of Campylobacter enteritis for the general population beyond the immunosuppressed or pregnant remains somewhat controversial, severe disease is not uncommon in undernourished children in the developing world, and the diminishing number of susceptible oral therapies is concerning. In the US and Europe, restricting the use of the fluoroquinolone enrofloxacin (Baytril) in poultry was a major regulatory step meant to limit the expansion of fluoroquinolones-resistant strains. However, nearly a decade later, fluoroquinolone resistance persists in the United States [66]. A recent analysis of feather meal used as animal feed strongly suggests that ongoing antimicrobial exposure may explain this [67]. Fluoroquinolone use in animals, particularly poultry for prophylaxis or growth promotion, is not regulated in most countries in the developing world, and increasingly affordable generics are allowing for expanded use, an issue that will need to be dealt with on an international scale.

CONCLUSION

Though progress has been made, the global burden of Campylobacter infection remains poorly understood. Advances in diagnostics as well as comprehensive and well-standardized prospective studies are helping to refine estimates of the acute and long-term burden of disease associated with a broad range of Campylobacter species. Better and more consistently applied assays are necessary for an improved understanding of the epidemiology of different Campylobacter species and to guide vaccine development. Antimicrobial resistance is of increasing concern and serves to underline the importance of control strategies as well as the regulation of antibiotic use.

KEY POINTS.

• Adoption of less-selective culture and culture-independent diagnostic tests suggests that both Campylobacter jejuni/coli and non-jejuni/coli infections have been frequently underdetected or misidentified.

• The burden of gastroenteritis as well as long-term consequences associated with Campylobacter infection in developing countries remains to be clearly defined.

• Significant progress has been made towards a Campylobacter vaccine, and epidemiologic studies of the importance of specific capsular serotypes are needed to help guide development.