Quantifying Transmission of Campylobacter jejuni in Commercial Broiler Flocks

Since meat from poultry colonized with Campylobacter spp. isa major cause of bacterial gastroenteritis, human exposure shouldbe reduced by, among other things, prevention of colonizationof broiler flocks. To obtain more insight into possible sourcesof introduction of Campylobacter into broiler flocks, it isessential to estimate the moment that the first bird in a flockis colonized. If the rate of transmission within a flock wereknown, such an estimate could be determined from the changein the prevalence of colonized birds in a flock over time. Theaim of this study was to determine the rate of transmissionof Campylobacter using field data gathered for 5 years for Australianbroiler flocks. We used unique sampling data for 42 Campylobacterjejuni-colonized flocks and estimated the transmission rate,which is defined as the number of secondary infections causedby one colonized bird per day. The estimate was 2.37 ±0.295 infections per infectious bird per day, which impliesthat in our study population colonized flocks consisting of20,000 broilers would have an increase in within-flock prevalenceto 95% within 4.4 to 7.2 days after colonization of the firstbroiler. Using Bayesian analysis, the moment of colonizationof the first bird in a flock was estimated to be from 21 daysof age onward in all flocks in the study. This study providesan important quantitative estimate of the rate of transmissionof Campylobacter in broiler flocks, which could be helpful infuture studies on the epidemiology of Campylobacter in the field.

INTRODUCTION

Top

INTRODUCTION

Campylobacter spp. are a common cause of diarrhea in humans,and many cases of campylobacteriosis are associated with thehandling and consumption of contaminated poultry meat (5). Studiesof the epidemiology of Campylobacter have resulted in the implementationof biosecurity and hygienic measures on poultry farms and slaughterhouseswith the ultimate goal of reducing human exposure (6, 16). Thesemeasures likely contributed to a reduction in the number ofCampylobacter-positive broiler flocks. Nevertheless, contaminatedmeat is still on the market (10), and a further reduction inthe prevalence of Campylobacter-positive flocks is considerednecessary by public health authorities in many countries (4).

Clearly, more knowledge concerning the mechanism of introductionof Campylobacter into a flock is essential for improving thecurrent control programs. This, in turn, requires an estimateof the moment that a flock becomes colonized. However, it doesnot seem feasible to detect the first bird that is colonizedin a commercial broiler flock because of flock size and thenecessary sampling frequency.

An alternative approach is to determine the transmission rate(β) of Campylobacter within a flock. β, which is definedas the number of secondary infections caused by one colonizedbird per day, determines the rate of increase in the numberof colonized birds over time. It can be used to determine themoment of introduction from field data for increasing Campylobacterprevalence over time. The estimates for β that have beenobtained in experimental studies (15, 17) are 1.04 to 1.13 perday. However, experimental conditions differ substantially fromthe field situation, which implies that the β in commercialflocks should also be estimated.

A series of field studies in Australia were carried out between1999 and 2004 in which broiler flocks were sampled daily toweekly. The aim of these studies was to develop an understandingof the epidemiology of Campylobacter in Australian broiler flocks.We analyzed data from the unique data set obtained in thesestudies to estimate the rate of Campylobacter transmission incommercial broiler flocks. Additionally, we estimated the momentthat the first bird in a flock was colonized with Campylobacter(for reasons of convenience, we refer to this event as the momentof introduction of Campylobacter into a flock) and assessedhow accurately moments of introduction can be estimated.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Data set.
Three longitudinal studies were carried out in southeast Queensland,Australia, between 1999 and 2004. A subset of the full dataset was selected for this study; for flocks to be included inthis study, they had to have been sampled at least twice, andat least one Campylobacter-positive dropping had to be detected.Flocks with migration barriers separating groups of birds withina shed and flocks whose sampling age was not stated were excluded.A total of 42 flocks met all the inclusion criteria (see thesupplemental material for details).

At each sampling age randomly selected individual fecal or cecaldroppings were collected immediately after the droppings wereproduced. Birds that produced these droppings were not marked,as the likelihood of sampling a single bird more than once aday was considered limited for the large populations used. Carewas taken to ensure that each sample was collected without anycontaminating material. The samples were collected with a sterileswab, which was immediately placed in a sterile container. Sampleswere kept on ice during transport to the laboratory and werestreaked on Karmali Campylobacter agar base (Oxoid CM935; Oxoid,Melbourne, Australia) containing Campylobacter selective supplement(Oxoid SR167E) immediately after arrival in the laboratory.Agar plates were incubated at 42°C for 48 h in an incubatorwith an atmosphere consisting of 85% N₂, 10% CO₂, and 5% O₂.Colony morphology and cell motility as determined by phase-contrastmicroscopy were used to confirm identification to the genuslevel. Single colonies from a number of positive samples fromeach positive flock were subcultured on sheep blood agar andincubated as described above before they were identified tothe species level by oxidase, catalase production, and hippuratehydrolysis tests.

The final data set consisted of data for sampling events foreach flock, including the age of the flock (t), sample size(n_t), flock size at the start of the rearing period (N), andnumber of positive samples (x_t) (see the supplemental materialfor details).

Modeling of Campylobacter transmission in broiler flocks.
It is generally assumed that following the onset of Campylobactercolonization of the gut, broilers shed these bacteria for therest of their lives (1). Therefore, we assumed that birds wereeither susceptible (noncolonized) or infectious (colonized)and that an increase in prevalence could be described by a susceptible-infectioustype of mathematical model (12, 17). In this model, susceptiblechicks can be colonized upon contact with an infectious chick,which occurs at rate βi(t), where i(t) is the proportionof infectious chicks in the shed. β is considered the meannumber of chickens that can be colonized by one infectious chickenper day in a susceptible population.

In the large populations of broilers usually present in commercialflocks, the change in i(t) can be approximated by the deterministicdifferential equation:

Formula 1 (1)
whose the solution is the logistic curve described by

Formula 2 (2)
in which ${tau}$ is the time at which50% of the birds are infected. From ${tau}$ , β, and N, the timewhen transmission starts (t₀) is calculated as follows:

Formula 3 (3)

Estimation of β in the study population.
Because the increase in i(t) follows a logistic curve, βis estimated by a logistic regression analysis of x_t. However,this is possible only if for at least two samples only someof the swabs are positive, because only then is the steepnessof the increase in prevalence (determined by β) observed.Eight flocks met this criterion. The susceptible-infectiousmodel was fitted for each of these eight flocks by logisticregression of the binomially distributed x_t (and n_t) with tas covariate and exp( ${tau}$ ) as the intercept. The model fits resultedin eight estimates for β and ${tau}$ , from which t₀ could be estimatedby using equation 3. Confidence intervals for β and t₀were derived by profile likelihood analysis (3), which was carriedout using Mathematica (version 6.0; Wolfram Research, Inc. [http://www.wolfram.com]).

Because flocks with low β have been more likely to meetthe inclusion criteria for the logistic regression analysisthan flocks with high β, bias would be introduced if theeight estimates were considered representative of all commercialbroiler flocks. Therefore, in the next step, we estimated themean (µ_β) and standard deviation ( ${sigma}$ _β) of a normaldistribution of β among all flocks by maximizing the likelihoodfunction

Formula 4 (4)
where p(β)is the density of the normal distribution of β, i(t) isthe logistic curve for equation 2, and t, x_t, and n_t are thedata for all flocks. This maximum likelihood estimation methoddoes not result in any flock-specific estimates for β ort₀; estimation results are limited to µ_β and ${sigma}$ _β,with 95% confidence intervals derived by profile likelihood(3). Although β cannot be negative, the assumption of anormal distribution is valid, because the distribution is wellabove zero (see Results). The analysis was carried out usingMathematica.

Estimation of the t₀ for the study population.
Because no flock-specific estimates of the moment of introduction(t₀) could be obtained with the analysis described above, weused Markov chain Monte Carlo integration to obtain Bayesianposterior distributions of β and ${tau}$ , and therefore t₀, foreach flock separately (7). The prior distribution of ${tau}$ was uninformative(flat), whereas the prior distribution of β was the normaldistribution resulting from maximum likelihood estimation (withµ_β = 2.37 day^–1 and ${sigma}$ _β = 0.295 day^–1).The likelihood function was

Formula 5 (5)
We used the maximum likelihood results for all 42 flocks asinput for estimation of t₀ in the same flocks, but it wouldhave been more correct to divide the data into two mutuallyexclusive subsets and use each subset only once. However, theapproach which we used resulted in more precise estimates ofµ_β and ${sigma}$ _β and estimates of t₀ for all flocksinstead of only a subset. The possible errors resulting fromour approach were minor, as determined by a separate sensitivityanalysis of estimation of t₀ to correctness of the prior distributionof β (see below).

The posterior distribution was sampled 10,000 times by single-componentMetropolis-Hastings sampling, after 100 samples were used forburn-in (8). The means of the (normal) proposal distributionsfor the (i + 1)th samples of both β and ${tau}$ were the ith samples.The standard deviation of the proposed distribution of βwas 0.295 day^–1; the standard deviation of the proposalfor ${tau}$ was determined from the data set, and it was one-quarterof the time interval between the last 0% prevalence sample andfirst 100% prevalence sample. This resulted in means and 95%credible intervals for t₀ for 40 of the 42 flocks (for 2 flocks,the flock size was not known). The sampling algorithm was programmedusing Mathematica.

Accuracy of t₀ estimation.
Although Bayesian estimation of t₀ does provide 95% credibleintervals, these intervals are based on a logistic curve whichdoes not take account of the stochastic nature of transmissionin the early phase of a Campylobacter outbreak. Therefore, weassessed the accuracy of the method by estimating t₀ using simulatedoutbreaks in which the estimates could be compared to the realvalue.

We simulated 10,000 outbreaks in flocks consisting of 20,000chicks with values for β sampled from the estimated distribution(µ_β = 2.37 day^–1; ${sigma}$ _β = 0.295 day^–1),with a t₀ of 0. Simulations were carried out as described previously(17) with three sample sizes (10, 20, and 100 birds) and threesampling frequencies (once every day, every third day, or everyseventh day). This resulted in 10,000 simulated data sets forall combinations of sample sizes and sample frequencies, whichwere subsequently analyzed by the Bayesian method describedabove to obtain posterior means for 1,000 samples of the posteriordistribution.

In addition to these 10,000 simulations, where the populationdistribution of β (used in the simulations) was identicalto the prior distribution in the Bayesian analysis, we assessedthe accuracy by simulation with other distributions, based onthe confidence intervals for µ_β (2.19 to 2.58 day^–1)and ${sigma}$ _β (0.144 to 0.488 day^–1). Four new data setswere simulated, including two data sets with different µ_β(2.19 and 2.58 day^–1, with ${sigma}$ _β kept at 0.295 day^–1)and two data sets with different ${sigma}$ _β (0.144 and 0.488 day^–1,with µ_β kept at 2.37 day^–1). This resultedin four sets of 1,000 simulated data sets for 10 samples every1, 3, or 7 days. These data were used to estimate t₀ as describedabove.

RESULTS

Top

RESULTS

Data set.
The flock age at the time of first detection of Campylobacter-positivedroppings varied from 24 to 54 days. For 14 of 42 flocks theprevalence was more than 95% at the time of first detection.For an additional 10 flocks in which the prevalence was up to10% at the time of first detection, sampling was continued untilthe flocks were found to be fully colonized (>95% prevalence),which took 4.4 days on average (range, 3 to 6 days) (see thesupplemental material for details).

For 38 of 42 flocks, all Campylobacter isolates were identifiedas Campylobacter jejuni. Both C. jejuni and Campylobacter coliwere found in four flocks. Because insufficient data concerningC. coli outbreaks were available, our analyses were based solelyon C. jejuni outbreaks.

Estimation of β and t₀ for the study population.
For eight flocks with sufficient data, flock-specific βand t₀ were estimated by logistic regression. The point estimatesfor β ranged from 1.3 to 3.1 day^–1 (Fig. 1), andt₀ ranged from 21 to 35 days (see the supplemental materialfor details).

View larger version (11K):
[in this window]
[in a new window]

FIG. 1. (Top) Eight estimates of Campylobacter β (with 95% confidence intervals), obtained by logistic regression. (Bottom) Estimated density of the normal distribution of β in the Australian study population, obtained by maximum likelihood estimation (µ_β = 2.37 day^–1; ${sigma}$ _β = 0.295 day^–1).

The estimates for the distribution of β, which were basedon 42 outbreaks, were as follows: µ_β = 2.37 day^–1(95% confidence interval, 2.19 to 2.58 day^–1) and ${sigma}$ _β= 0.295 (95% confidence interval, 0.144 to 0.488 day^–1)(Fig. 1). Figure 2 shows how fast flocks of 20,000 broilerswere colonized for the mean β, a low β, and a highβ within the distribution of β.

View larger version (19K):
[in this window]
[in a new window]

FIG. 2. Epidemic curves for three different values of the estimated distribution of β (lower limit, point estimate, and upper limit) and an experimentally derived β estimate (17) for flocks consisting of 20,000 broilers.

For 40 Campylobacter outbreaks in which the population sizewas known we estimated t₀ by Bayesian analysis. The results(Fig. 3) show that colonization did not take place before chickensreached an age of 3 weeks. On average, t₀ was 4.8 days (range,2.2 to 9.3 days) earlier than the first detection of Campylobacter-positivesamples (see the supplemental material for details).

View larger version (6K):
[in this window]
[in a new window]

FIG. 3. Histogram showing the t₀ for 40 Australian C. jejuni outbreaks.

The estimates obtained by logistic regression and Bayesian analysisdid not differ more than 0.6 day except for one flock (3.4-daydifference).

Accuracy of t₀ estimation.
The accuracy of the Bayesian method was assessed by analyzingsimulated outbreaks; the results are shown in Table 1. If theprior distribution used for the analysis reflected the underlyingβ in the population correctly, the t₀ was estimated within3 days even with weekly sampling of only 10 birds.

View this table:
[in this window]
[in a new window]

TABLE 1. Accuracy of Bayesian estimation (with prior distribution µ_β = 2.37 day^–1 and ${sigma}$ _β = 0.295 day^–1) of the posterior mean time of Campylobacter introduction into a flock of 20,000 broilers, with different sample sizes and sampling intervals

The Bayesian estimation method was not very sensitive to thecorrectness of the prior distribution (within the margins ofthe 95% confidence intervals), and the sensitivity was greatestif µ_β was lower or ${sigma}$ _β was higher, resulting ina margin of 3.3 days with weekly sampling of 10 birds.

DISCUSSION

Top

DISCUSSION

The first aim of this study was to determine the β of Campylobacterin commercial broiler flocks. β, based on the field data,was estimated to be 2.37 ± 0.295 day^–1, which meansthat one colonized bird could, on average, infect 2.37 birdsper day. This implies that in a flock of 20,000 broilers thewithin-flock prevalence of Campylobacter would increase froma prevalence of one colonized bird to a prevalence of 95% within1 week.

Previous studies showed that flocks determined to be Campylobacternegative at one moment could appear to be fully colonized within1 week, suggesting that Campylobacter was introduced in theintervening week (2, 9, 11). This seems consistent with ourfindings, but the sample size used in these studies did notallow detection of positive flocks at the start of an epidemic.This implies that bacteria could have been introduced at someprevious time.

The second aim of this study was to estimate the moment of colonizationof the first bird in a flock (t₀) by using the estimated β.The estimates of t₀ were all greater than a flock age of 21days. The estimate of t₀ was accurate within 3 days in 90% ofthe cases, even if only 10 birds were sampled weekly. The sensitivityto incorrectness of the prior distribution was low. The apparentlack of Campylobacter transmission in the first 3 weeks of life,which seemed to have occurred in our study, has been describedpreviously (11). It is difficult or probably impossible to determinewhether in the first 3 weeks no introduction occurred or whetherintroduction of bacteria did occur but did not result in colonizationof the birds. The risk of introduction of Campylobacter canbe high throughout the rearing period, but it could also behypothesized that this risk may increase over time. Anotherpossibility is that young chicks are less susceptible, as indicatedby Ringoir et al. (13) and Sahin et al. (14), who demonstratedthat age and maternally derived immunity had an effect on susceptibility.More insight into the underlying mechanism of this phenomenonmay provide clues for prevention of Campylobacter colonization.

Conclusions.
A unique data set describing the change in prevalence of Campylobacter-colonizedbirds in commercial broiler flocks was used to quantify C. jejunitransmission under field conditions. The estimated β impliesthat in a flock of 20,000 broilers the within-flock prevalenceof Campylobacter increases to 95% within 1 week after colonizationof the first bird. Since such rapid spread has not been describedpreviously on the basis of experimental transmission experiments,this study provides important new quantitative information onthe epidemiology of Campylobacter in broilers.

This study also showed how the β can be used to estimatewhen the first bird is colonized. We showed that this methodfor estimation was accurate, and therefore this method is apromising method for further studies of mechanisms of Campylobacterintroduction, because it allows focusing on the chronology ofevents.

Additionally, t₀ estimation can result in an accurate descriptionof the period in which no transmission of Campylobacter occurs,like the first 3 weeks for our study population. Consequently,interventions aimed at prevention of introduction of and subsequentcolonization by Campylobacter might better be targeted at thesecond half of the rearing period, which in our study populationcould be considered a high-risk period.

ACKNOWLEDGMENTS

This work was funded in part by the Australian funding bodyRIRDC.

FOOTNOTES

* Corresponding author. Mailing address: Department of Farm Animal Health, Faculty of Veterinary Medicine, Utrecht University, Yalelaan 7, 3584 CL Utrecht, The Netherlands. Phone: (31) 30 2531248. Fax: (31) 30 2521887. E-mail: tvgerwe{at}gmail.com

Published ahead of print on 1 December 2008.

Supplemental material for this article may be found at http://aem.asm.org/.

REFERENCES

Top