Although diarrheal illnesses are recognized as both a cause and effect of undernutrition, evidence for the effect of specific enteropathogens on early childhood growth remains limited. We estimated the effects of undernutrition as a risk factor for campylobacteriosis, as well as associations between symptomatic and asymptomatic Campylobacter infections and growth.
Using data from a prospective cohort of 442 children aged 0–72 months, the effect of nutritional status on the incidence of Campylobacter infection was estimated using uni- and multivariate Poisson models. Multivariate regression models were developed to evaluate the effect of Campylobacter infection on weight gain and linear growth. Overall, 8.3% of diarrheal episodes were associated with Campylobacter (crude incidence rate = 0.37 episodes/year) and 4.9% of quarterly asymptomatic samples were Campylobacter positive. In univariate models, the incidence of Campylobacter infection was marginally higher in stunted than non-stunted children (IRR 1.270, 95% CI (0.960, 1.681)(p = 0.095). When recent diarrheal burdens were included in the analysis, there was no difference in risk between stunted and unstunted children. Asymptomatic and symptomatic Campylobacter infections were associated with reduced weight gain over a three-month period (65.5 g (95% CI: −128.0, −3.0)(p = 0.040) and 43.9 g (95% CI:−87.6, −1.0)(p = 0.049) less weight gain, respectively). Symptomatic Campylobacter infections were only marginally associated with reduced linear growth over a nine month period (−0.059 cm per episode, 95% CI: −0.118, 0.001)(p = 0.054), however relatively severe episodes were associated with reduced linear growth (−0.169 cm/episode, 95% CI −0.310, −0.028)(p = 0.019).
Our findings suggest that Campylobacter is not as benign as commonly assumed, and that there is evidence to support expanding the indications for antibiotic therapy in campylobacteriosis in children.
Campylobacter is a common cause of diarrheal disease among children at risk for growth failure in the developing world. We evaluated risk factors for Campylobacter infection as well as the association between symptomatic and asymptomatic Campylobacter infections and child growth over three and nine-month periods. Undernourished (stunted) children were more likely to experience a Campylobacter infection, but adjusting for a recent history of diarrheal disease attenuated this relationship. Both symptomatic and asymptomatic infections were associated with poorer weight gain and symptomatic Campylobacter infections were marginally associated with poorer linear growth on an order similar to what has been reported for other bacterial pathogens and less than what has been reported for some protozoal and parasitic infections. In a subset of severe infections that made up about twenty percent of total illnesses, the associations were poorer growth were of greater magnitude. Campylobacter infections are frequently viewed as benign, but our study suggests that this is not always the case. Rapid diagnostics for Campylobacter jejuni and coli could attenuate acquired linear growth deficits in populations where campylobacteriosis is highly endemic by facilitating improved case management.
Citation: Lee G, Pan W, Peñataro Yori P, Paredes Olortegui M, Tilley D, Gregory M, et al. (2013) Symptomatic and Asymptomatic Campylobacter Infections Associated with Reduced Growth in Peruvian Children. PLoS Negl Trop Dis 7(1): e2036. doi:10.1371/journal.pntd.0002036
Editor: Joseph M. Vinetz, University of California San Diego School of Medicine, United States of America
Received: July 25, 2012; Accepted: December 12, 2012; Published: January 31, 2013
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: The study was funded by the National Institute of Health (K01-TW05717 to MK) and GL was supported by T32HD046405. The funding agency had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Campylobacter is a common cause of diarrhea among infants and children in the developing world . Risk factors include poor sanitation and close contact with animals . Among the known species and subspecies of Campylobacter, C. jejuni is the most frequently isolated, followed by C. coli . Campylobacter-associated diarrhea is generally acute but self-limiting . Unlike other common bacterial causes of diarrheal disease, it is frequently isolated from stools up to weeks following an episode, with reported mean excretion times following diarrhea as long as 14 days , .
Asymptomatic Campylobacter infections are common among older children, and are associated with prolonged excretion in a significant percentage of cases . In the developing world, asymptomatic Campylobacter carriage is more common among malnourished children; an observation that has been confirmed individually for C. jejuni , , and other species . This has led to the suggestion that in this context it is an opportunistic infection, perhaps related to undernutrition-induced immunosuppression . However, the relationship between infection and undernutrition is complex and the effect of Campylobacter infection and carriage on childhood growth has not to our knowledge previously been quantified using a longitudinal study design.
Biologically, there is reason to believe that Campylobacter infection may have an enduring impact on childhood growth. In addition to rarer immune-mediated long term sequelae including reactive arthritis and Guillain-Barré syndrome, Campylobacter is a risk factor for post-infectious inflammatory bowel syndrome and, more controversially, inflammatory bowel disease –. This is consistent with the hypothesis that there is an adverse physiologic insult extending beyond the period of acute diarrhea, at least in a subset of cases. Campylobacter infection has been found to affect epithelial barrier integrity , suggesting that prolonged excretion may be associated with persistent mucosal injury.
Our objectives were to evaluate undernutrition as a risk factor for Campylobacter infection and to estimate the effects of symptomatic and asymptomatic Campylobacter infections on early childhood growth. Longitudinal analysis methods were used to disentangle the potentially bi-directional relationship between undernutrition and Campylobacter in a population with high rates of diarrheal disease and chronic undernutrition .
Data were from an age-stratified, prospective, community-based study of 442 children 0–72 months of age in a semi-rural community in the Peruvian Amazon, between 2002 and 2006 . The cohort and study design were described previously , . Briefly, recruitment, socioeconomic and demographic information were based on community censuses conducted before and during the study period. Every third child in a list of age-eligible children living in the study community without another family member enrolled was invited to participate. Therefore, only one child per household was enrolled in the study at any given time. Enrollment was continuous through the study period. The overall objective was to compare the associations between common etiologies of diarrhea and early childhood growth , , with the sample size based on this objective.
Length (children 0–23 months) or height (children 24–72 months) were measured using a marked platform with a sliding footboard, and weight was measured using Salter scales (Salter Housewares Ltd, Tonbridge, England). This occurred monthly on the day of their birth. For analyses, the measures were converted to z-scores (height-for-age-Z (HAZ) and weight-for-height-Z (WHZ)) using the 2006 WHO standards .
Participating families were visited three times weekly by a trained health promoter to document the number and consistency of stools passed by the child, as well as other symptoms like fever, and reported medication use, including specific antibiotics. This generated a continuous history of diarrheal disease for each child.
Diarrhea was defined by three or more semi-liquid stools reported over a 24-hour period, with episodes separated by at least three symptom-free days. Stool samples were collected as soon as possible after the case definition was met, and not more than two days after the episode ended. One sample was sought for all episodes; however, children who were culture-negative for Campylobacter and Shigella but continued to have diarrhea provided a sample every fourth day until the episode ended. Asymptomatic stool samples were collected quarterly to detect background enteropathogen carriage.
Fresh stool samples were placed in glycerol buffered saline and Cary-Blair medium and transported in a cooler from the field site for same-day plating. Briefly, one gram of feces in one milliliter glycerol buffered saline, pH 7·2, was received from the field and agitated briefly. A 100 ul aliquot was then placed on a 0.45 um sterile nitrocellulose filter overlying a Columbia blood agar plate and allowed to diffuse passively for 30 minutes. The filter was removed with sterile forceps and the plate was incubated in a microaerophilic atmosphere (5% O2, 10% CO2, 85% N) at 42 C for 48 hours. Suspected colonies were evaluated by gram staining, oxidase and catalase testing. Colonies confirmed as Campylobacter by these methods were then evaluated by the hippurate hydrolysis tube test to distinguish Campylobacter jejuni from Campylobacter coli The specimen aliquot in Cary-Blair was plated on MaConkey, Hektoen, XLD, and TCBS agar plates as previously reported . Additionally, for all samples, technician-observed gross blood and mucus were noted. Methylene blue was used to identify fecal leukocytes, and occult blood was identified by Hemoccult fecal occult blood test (SmithKline Diagnostics, Inc., Sunnyvale, California).
In cases where the child was still symptomatic when bacteriological culture confirmed a positive Campylobacter (or other enteropathogen) result, appropriate antibiotic therapy was administered (azithromycin or erythromycin). In the case of Campylobacter, a 72-hour delay between specimen acquisition and microbiologic reporting meant that the majority of cases resolved before a result was received.
Written informed consent to participate in the study was provided by all subjects. Parents consented on behalf of their children. The study protocol was approved by the Institutional Review Boards of Johns Hopkins Bloomberg School of Public Health (Baltimore, MD), US Navy Medical Research Center (Silver Spring, MD), Asociacion Benefica PRISMA (Lima, Peru), and the Regional Health Department of Loreto Peru.
Definitions and Descriptive Analyses
Campylobacter infection was defined as asympyomatic if there was not a diarrheal episode associated with Campylobacter in the two weeks period preceding its culture identification. Symptomatic campylobacteriosis was discounted if another Campylobacter-episode had been detected less than 15 days prior (i.e. it was taken to be a persistent infection).
Baseline characteristics of children who experienced at least one case of Campylobacter-associated diarrhea during the course of the study were compared to children who did not experience any Campylobacter-associated diarrhea during the study period. Two-sided t-tests were used to compare the child's age of entry into the study and total time in the study, and multivariate poisson models adjusting for age were used to describe the incidence of non-campylobacter diarrhea. Running-average smoothed plots of incidence versus age, and the percentage of stools positive by etiology, were created (see Figure 1). Throughout the analyses, p-values of between 0.10 and 0.05 were considered to be “marginally significant”, p-values of less than 0.05 “significant”, and p-values of less than 0.01 “very significant”.
The peak incidence of Campylobacter-associated diarrhea occurs at approximately 18 months of age, and declines rapidly thereafter. However, its isolation rate in diarrheal and asymptomatic stool samples remains roughly constant from 18–72 months of age.
Risk Factors for Campylobacter Infection
Bivariate Poisson models with a child-level random intercept to account for host-specific susceptibility were built to examine potential risk factors for campylobacteriosis and asymptomatic Campylobacter. Risk factors signficant at the p<0.15 level in univariate models were included in a multivariate model. Sex, age (using linear splines with knots at 18 and 36 months), non-Campylobacter diarrheal incidence in the past 3 months, campylobacteriosis in the past three months, seasonal effects (as sine/cosine terms), SES (as log per-capita income), maternal education, maternal age, the the presence of a household water connection, a private household latrine, poultry in the household, breast-feeding status, and birth weight, were considered for inclusion. The effects of prior undernutrition on the risk of campylobacteriosis was tested by considering several variables related to anthropometric status as measured in the month prior to the detection of Campylobacter. We considered HAZ (height-for-age-Z) and WHZ (weight-for-height-Z) as continuous independent variables, as well stunting as a binary variable (HAZ <−2), and WHZ categorized as less than −1, between −1 and 0, and greater than 0 (WHZ less than −2 was rare in this cohort). In final models, stunting as a binary variable and WHZ as a categorical variable were used. Correlations and kappa statistics comparing prior nutritional status and prior diarrheal disease were calculated.
Effects of Campylobacter Infection on Growth
As it was assumed that deficits in weight gain would precede linear deficits, models were constructed to examine the effects of symptomatic and asymptomatic Campylobacter on weight over three month periods. Linear growth periods ranging from two to nine months were used to examine the effects of campylobacteriosis and asymptomatic Campylobacter infections on length/height. The longer interval was retained in the final models as we posited that durable linear growth deficits were of increased importance relative to short term deficits.
All models included a random intercept for each study child. The covariance structure was fixed such that the residuals took a third-order moving average form for change-in-weight models, and a first-order autoregressive form for change-in-height models. This choice was based on a visual inspection of the autocorrelation function as well as model fit. Seasonal effects were considered by including the terms and , where d is the day of the year and t is 365 , . Per-capita income , prior nutritional status (stunting and WHZ category as defined above), sex, birth weight, maternal age and education, and breast-feeding status, were also considered for inclusion. Variables were selected into the final model based on model fit.
The final models are shown in Equations 1 and 2. Two sets of models were constructed. The first estimates the association of symptomatic and asymptomatic Campylobacter on change in weight, after adjusting for stunting, WHZ category, birth weight, season, age and per-capita income (Equation 1). The second estimates the association of Campylobacter infection on length/height, controlling for the same factors (Equation 2).(1)In Equation 1, is a binary variable set to one if asymptomatic Campylobacter was encountered, and zero otherwise. Similarly, is one if campylobacteriosis is encountered (two cases of more than one symptomatic Campylobacter in a three-month period were re-classified to 1), and zero otherwise. is a continuous variable representing the number of diarrheal episodes in the three months unassociated with Campylobacter. In weight models, anthropometry that was measured during or within two days of a Campylobacter associated diarrheal episode was excluded from the analysis, in order to reduce the possibility that reductions in weight were related to clinical or subclinical dehydration.(2)In Equation 2, is a continuous variable taking values between one and three, the number of quarters in the nine month interval in which asymptomatic Campylobacter infection was observed. and are continuous variables representing the total incidence of Campylobacter-associated and non- Campylobacter-associated episodes of diarrhea in the interval.
Effects of Treated Versus Untreated Campylobacteriosis on Growth
A second set of models were created, identical to those described above, except that symptomatic Campylobacter infection was subdivided into two types: treated, and untreated, which were tested separately. Therefore, the incidence of treated and untreated infections were taken as two separate binary variables (in weight models), or two separate continous variables (in height models).
All analysis was performed using Stata 11 (Statacorp, College Station, TX).
Ninety-six percent of families invited to participate in the study consented (five families refused). Of those enrolled, 20.8% (N = 92) dropped out during the study (median time in study = 15 month); permanent migration outside the community was the principal reason. 442 children were enrolled, nine (2.0%) were excluded because of limited anthropometric data, defined as fewer than three consecutive months of anthropometry. In total, we conducted 838 total child-years of surveillance and collected anthropometric data during 10,985 study visits. Child characteristics are outlined in Table 1.
There were 3973 diarrheal episodes observed during the surveillence period. In 3667 (92.3%) of these, a stool sample was collected and tested in association with the episode. Campylobacter was isolated in 306 (8.3%). There was a low rate of co-infections between Campylobacter species (e.g. C. jejuni+C. coli) (Table 2), a finding in agreement with prior reports .
Nearly nineteen percent (18.9%) of campylobacteriosis episodes were treated with appropriate antibiotic therapy. Treated episodes had a greater duration than untreated infections (8.1 versus 2.8 days, p<0.01), and were more frequently associated with mucus (57.7% versus 33.7% of episodes, p<0.01); however, no differences were observed in the presence of gross blood (13.6% versus 7.9% p = 0.18), reported fever (42.4% versus 33.3%, p = 0.19), or the presence of fecal leucocytes or occult blood.
No significant difference were detected in fecal leucocytes, occult blood, or technician-observed mucus between asymptomatic stools with or without Campylobacter (6.9% v. 6.2%, 9.2% v. 8.2%, 6.3% v 4.0%, respectively). However gross blood was higher in Campylobacter-positive asymptomatic stools (1.7 versus 0.5%, p = 0.03).
In contrast, fecal leucocytes, occult blood, and gross blood were more common in Campylobacter-positive, as compared to Campylobacter-negative diarrheal samples (12.9 versus 8.0%, p<0.01, 17.0% versus 11.1% p<0.01, and 9.0% versus 4.4, p<0.01, respectively). However, no differences were detected for technician-observed mucus (38.3% versus 34.8%, p = 0.214).
The incidence of Campylobacter-associated diarrhea peaked at 0.7 episodes/child-year, in 18 month old children (Figure 1). There was then a steep decline in the incidence of Campylobacter diarrhea, the rate of isolation of Campylobacter as a proportion of diarrheal and asymptomatic declined more gradually (Figure 1).
Risk Factors for Campylobacter Infection
Nearly fifty percent (49.7%) of children had a Campylobacter infection during the study period. Ninety-two (21.2%) experienced campylobacteriosis but no asymptomatic Campylobacter; 47 experienced asymptomatic Campylobacter only, and 79 experienced symptomatic and asymptomatic infections. In addition to the other risk factors described below, children who experienced Campylobacter infection during the course of the study had a higher incidence of diarrhea overall (10.9 versus 7.6 episodes/yr among children who ever versus never had Campylobacter); entered the study at a younger ages (mean 19.5 versus 31.8 months), and remained in the study longer (mean 34.7 versus 22.8 months in study).
Both asymptomatic and symptomatic Campylobacter infections were associated with younger age, a recent history of diarrheal illness and Campylobacter diarrhea, and less maternal education (Table 3). Per-capita income and the presence of a household water connection were predictive of Campylobacter diarrhea, but not asymptomatic infection (Table 3). Child sex, maternal age, breastfeeding status, seasonal effects, the presence or a private family latrine, birthweight, and the presence of poultry in the household, were not found to predict either symptomatic Campylobacter incidence or asymptomatic Campylobacter, and per-capita income and the presence of a household water connection were found to predict Campylobacter diarrhea, but not asymptomatic Campylobacter infection (Table 2).
Stunting was a significant risk factor for Campylobacter diarrhea when recent diarrhea was not also included in the model. When recent diarrhea was included, the association of stunting with campylobacter risk decreased from an IRR of 1.270 compared to non-stunted children, to an IRR of 1.175. Stunting and recent diarrhea were not strongly correlated (rho = 0.050).
Effects of Campylobacter Infection on Growth
Asymptomatic Campylobacter and Campylobacter-associated diarrhea were each associated with reduced weight gain over a three-month period (65.5 and 43.9 grams less weight gain, respectively) (Table 4).
Symptomatic Campylobacter, but not asymptomatic Campylobacter infection, was marginally associated with reduced linear growth over nine-month intervals (Table 5). Each symptomatic Campylobacter episode was marginally associated with 0.059 cm less linear growth, but this was not significant at the p<0.05 level (p = 0.054).
When subdivided into treated episodes (in which the child was still symptomatic at the time of a bacteriological confirmation of Campylobacter) and untreated episodes (in which the episode resolved before bacteriological confirmation), most of the impact of Campylobacter on weight gain was clustered among the more severe treated episodes (167.2 grams versus 11.0 grams deficit, p-value for difference <0.01). Similarly, the effect of treated Campylobacter on linear growth was five times greater than that of untreated symptomatic Campylobacter (−0.169 versus −0.034 cm deficit, p-value for difference = 0.091).
Both symptomatic and asymptomatic Campylobacter were associated with reduced weight gain over three-month periods. Additionally, symptomatic Campylobacter was marginally associated with reduced linear growth over nine-month periods. More severe Campylobacter episodes were associated with greater deficits in both weight gain and linear growth.
Because severe episodes were more likely to be treated, treatment was used to define severity in this analysis. However, it is also a confounding factor in estimating the true effect of severe campylobacteriosis on growth. Treatments were effective; 80.8% of isolates were susceptible to erythromycin, and 91.9% to azithromycin. Episodes ended on average 1.45 days after initiation of antibiotic therapy, which is 1.32 days earlier than episodes of comparable prior duration which remained untreated (data not shown). This is similar to other published reports .
Fifty-one percent of Campylobacter-associated diarrheal episodes were related to C. jejuni, and 30.4 percent to C. coli. Our decision to pool these species during the analysis was based on the limited number of samples available from each, and preliminary explorations of the data did not suggest any differences in association between C.jejuni and C.coli infections and growth. We cannot draw conclusions about differential effects of Campylobacter by species.
The association between co-infections and growth faltering is an interesting area of current research. We considered whether helminth infections might also have influenced growth. As previously reported, children were dewormed very frequently in this community, both due to own study protocols and parental initiative. Perhaps as a result, we found little association between helminth infections and growth . The number of instances where Campylobacter and a helminth were identified in the same stool was low, and we were unable to relate recent helminth infection (e.g. detected in the 0–3 months prior) and Campylobacter infections for the same reason, almost all detected helminths were treated very promptly. As a result, this particular co-infection is unlikely to have confounded the Campylobacter-growth relationship, and we are also unable to determine whether such co-infections have a greater association with growth faltering than either infection alone.
Our analysis looked at the impact of Campylobacter infection on growth velocity (that is, the change in growth over 3 and 9 month periods) rather than a child's attained weight or height, or absolute risk of stunting. This approached was necessitated by the data available; the cohort was not a birth cohort, and anthropometry and surveillance histories were not available from birth for most children. However, because growth is naturally a non-linear process, natural variability from period to period may attenuate the effects of diarrheal illness over short periods . Having said that, the modeling the effects of infection on short-term growth velocities remains worthwhile, especially in cohorts where anthropometric indices clearly indicate progressively growth faltering.
While there is good evidence that diarrhea as a non-specific illness, leads to reduced child growth and poorer nutritional status , there is a paucity of evidence associating specific enteropathogens with reduced growth. Protozoal infections, including Cryptosporidium and E. histolytica, have been found to have large negative impacts –, with a first cryptosporidium infection associated with 0.95 cm shorter mean height one year after the infection . Among enteric bacterial pathogens, ETEC-diarrhea has been shown to negatively impact weight gain, and Shigella to negatively impact linear growth, with an effect size of 0.073 cm per percent of days of Shigella-associated diarrhea, over a year . Our finding that an episode of Campylobacter, of average duration 3.8 days, was marginally associated with nine-month deficits in height in the range of 0.059 cm, suggest effects similar in magnitude to other bacterial infections. In this cohort, Campylobacter was isolated in diarrheal stools as frequently as Shigella, and slightly less frequently than Enterotoxigenic E. coli (approximately 8.3%, 8.2%, and 10.1% of diarrheal episodes respectively).
When examining our data in a non-longitudinal (i.e. cross sectional) manner, we observed more frequent Campylobacter infections among children with lower HAZ and WHZ, in agreement with prior reports , . However, in longitudinal models, prior WHZ failed to be predictive of campylobacteriosis, and prior stunting was only predictive of campylobacteriosis when recent diarrheal disease was not adjusted for in the analysis. In this cohort, stunting was common, while wasting, underweight, or severe undernutrition in any form, were rare. As the risk of diarrheal disease may be greatest among severely malnourished children, our population may be less than ideally suited to disentangling this effect. Similarly, younger infants (less than two years) were relatively underrepresented in this cohort, which limited our ability to test whether the association between Campylobacter and growth was mediated by age.
Recently, Gupta and colleagues reported that samples from a malnourished child found enteric pathogens in the family Campylobacteraceae 35-fold higher than in those of a well-nourished child . This finding supports theories that Campylobacter carriage may be related to some underlying pathology, such as perturbed gut microbial populations, preexisting intestinal injury, or undernutrition-related immunosuppression . Our finding that asymptomatic Campylobacter infection was associated with reduced weight gain suggests it may be a signal that a child is at risk of entering into a negative cycle of an accelerating incidence of diarrhea with the predicable negative impacts on linear growth.
It is often advised that diarrhea caused by Campylobacter does not warrant treatment, unless the patient is pregnant or HIV+ or the episode is associated with dysentery, fever or other severe symptoms . Our finding that Campylobacter is associated with deficits in growth is an argument for broader treatment among children under the age of five, as well as a justification for the development of field-ready rapid diagnostics for early Campylobacter jejuni and coli.
The magnitude of the linear growth deficit attributable to Campylobacter infection is modest. However, acquired growth deficits in early childhood are associated with a variety of long-term negative outcomes including poorer cognitive development , lower adult work capacity and income , and poorer pregnancy outcomes , making growth deficits an accepted surrogate of human potential lost. Our findings reinforce the importance of controlling Campylobacter infection. Prioritization of interventions and strategies against campylobacteriosis, such as the development of rapid diagnostics for early detection at the point of care, and the development of vaccines for prevalent Campylobacter serotypes, may improve child growth in the developing world.
We would like to thank Matilda Bustos Aricara, Victora Lopez Manuyama, Marla Judith Aricari Huanari, and Lleny Amasifuen Llerena for their hard work and thoughtful contributions in the field. Also, we would like to thank Dr. Laura Caulfield for her review of the manuscript and support and Juan Perez Bao for assistance with data management.
Conceived and designed the experiments: MK PPY MPO. Performed the experiments: CBC RB DT MG. Analyzed the data: GL WP MK RO. Contributed reagents/materials/analysis tools: DT MG. Wrote the paper: GL MK.
- 1. Oberhelman R, Taylor DN (2000) Campylobacter infections in developing countries. In: Nachakin I, Blaser MJ, editors. Campylobacter. Washington: American Society for Microbiology. pp. 139–153.
- 2. Amieva MR (2005) Important bacterial gastrointestinal pathogens in children: a pathogenesis perspective. Pediatr Clin N Am 52: 749–777 doi:10.1016/j.pcl.2005.03.002.
- 3. Janssen R, Krogfelt KA, Cawthraw SA, Van Pelt W, Wagenaar JA, et al. (2008) Host-pathogen interactions in Campylobacter infections: the host perspective. Clin Microbiol Rev 21: 505–518 doi:10.1128/CMR.00055-07.
- 4. Figueroa G, Galeno H, Troncoso M, Toledo S, Soto V (1989) Prospective study of Campylobacter jejuni infection in Chilean infants evaluated by culture and serology. Infect Immun 27: 1040–1044.
- 5. Rao MR, Naficy AB, Savarino SJ, Abu-elyazeed R, Wierzba TF, et al. (2001) Pathogenicity and convalescent excretion of Campylobacter in rural Egyptian children. Am J Epidemiol 154: 166–173. doi: 10.1093/aje/154.2.166
- 6. Reed RP, Friedland IR, Wegerhoff FO, Khoosal M (1996) Campylobacter bacteremia in children. Pediatr Infect Dis J 15: 345–348. doi: 10.1097/00006454-199604000-00012
- 7. Da Silva Quetz J, Nunes Lima IF, Havt A, Bobo de Carvalho E, Leal Lima N, et al. (2010) Campylobacter jejuni and Campylobacter coli in children from communities in Northeastern Brazil: molecular detection and relation to nutritional status. Diagn Micr Infec Dis 67: 220–227 doi:10.1016/j.diagmicrobio.2010.02.025.
- 8. Fernández H, Vera F, Villanueva MP, García A (2008) Occurrence of Campylobacter species in healthy well-nourished and malnourished children. Braz J Microbiol 39: 56–58 doi:10.1590/S1517-83822008000100013.
- 9. Gradel KO, Nielsen HL, Schønheyder HC, Ejlertsen T, Kristensen B, et al. (2009) Increased short- and long-term risk of inflammatory bowel disease. Gastroenterology 137: 495–501 doi:10.1053/j.gastro.2009.04.001.
- 10. Kirkpatrick BD, Tribble DR (2011) Update on human Campylobacter jejuni infections. Current Opin Gastroen 27: 1–7 doi:10.1097/MOG.0b013e3283413763.
- 11. Kalischuk LD, Buret AG (2010) A role for Campylobacter jejuni-induced enteritis in inflammatory bowel disease? Am J Physiol Gastrointest Liver Physiol 298: G1–G9 doi:10.1152/ajpgi.00193.2009.
- 12. Ruigómez A, García Rodríguez LA, Panés J (2007) Risk of Irritable Bowel Syndrome After an Episode of Bacterial Gastroenteritis in General Practice: Influence of Comorbidities. Clin Gastroenterol H 5: 465–469 doi:10.1016/j.cgh.2007.02.008.
- 13. Chen ML, Ge Z, Fox JG, Schauer DB (2006) Disruption of tight junctions and induction of proinflammatory cytokine responses in colonic epithelial cells by Campylobacter jejuni. Infect Immun 74: 6581–6589 doi:10.1128/IAI.00958-06.
- 14. Kosek M, Peñataro Yori P, Pan WK, Paredes Olortegui M, Gilman RH, et al. (2008) Epidemiology of highly endemic multiply antibiotic-resistant shigellosis in children in the Peruvian Amazon. Pediatrics 122: e541–9 doi:10.1542/peds.2008-0458.
- 15. Lee G, Peñataro Yori P, Paredes Olortegui M, Pan W, Caulfield L, et al. (2012) Comparative effects of vivax malaria, fever and diarrhoea on child growth. Int J Epidemiol [Epub ahead of print]. doi:10.1093/ije/dyr190. doi: 10.1093/ije/dyr190
- 16. WHO (n.d.) Child Growth Standards (2006) Methods and development: Length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age.
- 17. Stolwijk A, Straatman H, Zielhuis G (1999) Studying seasonality by using sine and cosine functions in regression analysis. J Epidemiol Commun H 53: 235–238. doi: 10.1136/jech.53.4.235
- 18. Brown KH, Black RE, Becker S (1982) Seasonal changes in nutritional status and the prevalence of malnutrition in a longitudinal study of young children in rural Bangladesh. Am J Clin 36: 303–313.
- 19. Desai P, Standard K, Miall W (1970) Socio-economic and cultural influences on child growth in rural Jamaica. J Biosoc Sci 2: 133–143. doi: 10.1017/s0021932000007574
- 20. Royston P, Ambler G, Sauerbrei W (1999) The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol 28: 964. doi: 10.1093/ije/28.5.964
- 21. Ternhag A, Asikainen T, Giesecke J, Ekdahl K (2007) A Meta-Analysis on the Effects of Antibiotic Treatment on Duration of Symptoms Caused by Infection with Campylobacter Species. 44: 696–700 doi:10.1086/509924.
- 22. Karlberg J, Gelander L, Albertsson-Wikland K (1993) Distinctions between short-and long-term human growth studies. Acta Paediatr 82: 631. doi: 10.1111/j.1651-2227.1993.tb18029.x
- 23. Checkley W, Epstein LD, Gilman RH, Cabrera L, Black RE (2003) Effects of acute diarrhea on linear growth in Peruvian children. Am J Epidemiol 157: 166 doi:10.1093/aje/kwf179.
- 24. Mondal D, Petri WA, Sack RB, Kirkpatrick BD, Haque R (2006) Entamoeba histolytica-associated diarrheal illness is negatively associated with the growth of preschool children: evidence from a prospective study. T Roy Soc Trop Med H 100: 1032–1038 doi:10.1016/j.trstmh.2005.12.012.
- 25. Checkley W, Epstein L, Gilman RH, Black RE, Cabrera L, et al. (1998) Effects of Cryptosporidium parvum infection in Peruvian children: growth faltering and subsequent catch-up growth. Am J Epidemiol 148: 497–506. doi: 10.1093/oxfordjournals.aje.a009675
- 26. Mondal D, Minak J, Alam M, Liu Y, Dai J, et al. (2012) Contribution of enteric infection, altered intestinal barrier function, and maternal malnutrition to infant malnutrition in Bangladesh. Clin Infect Dis 54: 185–192 doi:10.1093/cid/cir807.
- 27. Black RE, Brown KH, Becker S (1984) Effects of diarrhea associated with specific enteropathogens on the growth of children in rural Bangladesh. Pediatrics 73: 799.
- 28. Gupta SS, Mohammed MH, Ghosh TS, Kanungo S, Nair GB, et al. (2011) Metagenome of the gut of a malnourished child. Gut Pathog 3: 9 doi:10.1186/1757-4749-3-7.
- 29. Grantham-McGregor S (1995) A review of studies of the effect of severe malnutrition on mental development. J Nutr 125: 2233.
- 30. Haas JD, Murdoch S, Rivera J, Martorell R (1996) Early nutrition and later physical work capacity. Nutr Rev 54: S41. doi: 10.1111/j.1753-4887.1996.tb03869.x
- 31. Black RE, Allen LH, Bhutta ZA, Caulfield LE, De Onis M, et al. (2008) Maternal and child undernutrition: global and regional exposures and health consequences. Lancet 371: 243–260. doi: 10.1016/s0140-6736(07)61690-0