Significance of interactions between Escherichia coli and respiratory pathogens in layer hen flocks suffering from colibacillosis-associated mortality

Abstract

This study aimed to examine the significance of interactions between Escherichia coli and various respiratory pathogens during outbreaks of colibacillosis-associated mortality in layer hen flocks under field conditions. For this purpose, a case–control study involving 20 control flocks with baseline mortality and 20 flocks with increased mortality due to E. coli septicaemia and polyserositis, was conducted. In each colibacillosis flock, blood samples were taken from 20 hens at the onset of clinical disease and three times thereafter at 2-week intervals. Control flocks of comparable ages were sampled in the same way. Pooled sera, taken at the first and last sampling, were examined for antibody titres against infectious bronchitis virus (IBV) and Newcastle disease virus (NDV), and the individual sera from all four samplings were examined for the presence and/or titres of antibodies against avian pneumovirus (APV), Mycoplasma gallisepticum, Mycoplasma synoviae and Ornithobacterium rhinotracheale. Titre increases were seen for IBV D274 (one control flock) and O. rhinotracheale (one control and one colibacillosis flock). An increase in per cent reactors was seen for APV (one control flock), and for M. synoviae (one control and two colibacillosis flocks). The study failed to detect any consistent interactions between E. coli and the aforementioned pathogens.

These results indicate that, at least as observed in this study, outbreaks of increased mortality resulting from colibacillosis are not necessarily associated with IBV, NDV, APV, M. gallisepticum, M. synoviae or O. rhinotracheale infections.

Introduction

Colibacillosis, caused by avian pathogenic Escherichia coli, is one of the main causes of economic loss in the poultry industry worldwide (Morris, 1989; Dho-Moulin, 1993; Yogaratnam, 1995; Elfadil et al., 1996; Barnes & Gross, 1997). Traditionally, septicaemia and polyserositis in laying hens represented only a limited part of the poultry problems related to colibacillosis. Although not proven by sound data it is generally accepted in the field that, from the mid-1990S onwards, the incidence of colibacillosis in layer flocks has substantially increased in many European countries. On the basis of field observations, Vandekerchove et al. 2004) described the current problems of colibacillosis outbreaks in laying hens as a specific entity characterized by acute mortality without prior clinical signs of disease, with septicaemia and polyserositis lesions at necropsy, and without a significant impact on egg production or egg quality. Most outbreaks occur around the period of peak production, and cumulative mortality may exceed 9%. In the majority of the flocks, the E. coli strains isolated belonged almost exclusively to the O78 serotype, and they usually possessed F11 fimbriae and flagella (Vandekerchove et al., in press).

In other types of poultry such as broilers, colibacillosis is often considered as a problem that occurs secondarily to other pathogens and/or unfavourable environmental conditions (Barnes & Gross, 1997). For layer hens, however, this aspect of the pathogenesis has not been clarified.

It was the purpose of the present study to examine whether infections with infectious bronchitis virus (IBV), Newcastle disease virus (NDV), avian pneumovirus (APV), Mycoplasma gallisepticum, Mycoplasma synoviae and Ornithobacterium rhinotracheale might play a role in outbreaks of colibacillosis in laying hens.

Materials and methods

Study design and flock history

The study was set up as a case–control design in 40 commercial caged layer flocks originating from 25 different farms in the northern part of Belgium, including 20 flocks suffering from colibacillosis-associated mortality and 20 clinically healthy control flocks of the same ages. On 14 farms one flock was sampled, while on one, two and eight farms four, three and two flocks were sampled, respectively. On four farms, both a colibacillosis and a control flock were sampled. All farms with more than one flock were multi-age farms. Ages of the control flocks varied between 21 and 74 weeks, and between 21 and 87 weeks in the flocks with a colibacillosis outbreak. Flock sizes ranged from 11 005 to 50 000. Breeds were Isa Brown (27/40), Lohman Selected Leghorn (6/40), Lohman Brown (3/40), Hisex Brown (3/40) or Hisex White (1/40). Maximum weekly mortality in the control flocks ranged from 0.07 to 0.30%, with an average of 0.13%; these mortalities did not have the typical characteristics of colibacillosis as described by Vandekerchove et al. (in press). In the colibacillosis flocks, they ranged from 0.26 to 1.71%, with an average of 0.72%. Increased mortality in the colibacillosis flocks lasted from 3 to 10 weeks, with a threefold to eightfold increase in the weekly number of dead hens within a 1-week to 3-week period following the onset of the outbreak. A decrease in egg production exceeding 2% was observed in one control and six colibacillosis flocks. A decrease in egg quality (percentage second-class eggs exceeding 1.5%) was seen in three colibacillosis flocks. In only one colibacillosis flock were both a decrease in egg production and egg quality seen (Vandekerchove et al., in press).

The 20 colibacillosis-affected flocks were identified on the basis of clinical characteristics, necropsy findings and the isolation of E. coli in abundant or pure cultures from the lesions according to the criteria of Vandekerchove et al. (in press).

Flocks suffering from an outbreak of colibacillosis were visited within 2 weeks of the onset of the increased mortality and three times thereafter at 2-week intervals. On each occasion serum was collected from the same 20 hens in five marked cages. A maximum of one cage per battery was sampled, alternately situated near the entrance of the house, at the far end of the house, and in between, to have layers representative of the entire flock. Control flocks of comparable ages without increased mortality were sampled in the same way.

Serology

All sera were examined for the presence of antibodies or antibody titres against IBV, NDV, APV, M. gallisepticum, M. synoviae and O. rhinotracheale.

The presence of antibodies against M. gallisepticum was determined in fresh, undiluted sera by a rapid slide agglutination test using Nobilis M. gallisepticum antigen (Intervet International, Boxmeer, The Netherlands). Samples with positive or doubtful reactions were retested with the Svanovir MG-Ab enzyme-linked immunosorbent assay (ELISA) (Svanova Biotech AB, Uppsala, Sweden). All other tests were performed on sera kept frozen until used.

Sera obtained from the first farm visits and those taken 6 weeks later were examined for their antibody titres against IBV and NDV. For this purpose, the 20 sera of each sampling were pooled and tested with haemagglutination inhibition (HI) using 8 haemagglutination (HA) units IBV M41, D274 and D1466, or 4 HA units IBV 4/91 as antigens for IBV, and 8 HA units La Sota NDV. A titre increase of more than two log₂ units in the pooled sera was considered as a seroconversion. If the titre increase amounted to two log₂ units, the corresponding individual sera were examined, in order to exclude the possibility of overestimation, (e.g. a recorded titre of 8 may represent an actual titre between 8.1 and 8.9). Only if the resulting titres yielded an average increase of two or more log₂ units was the seroconversion confirmed.

For APV, M. synoviae and O. rhinotracheale antibody detection, all 3200 sera obtained during the four successive samplings of each flock were tested individually. For APV and M. synoviae, the Svanovir APV-Ab ELISA (Svanova Biotech AB) and the Idexx MS (Idexx, Schiphol-Rijk, The Netherlands) ELISA were used, respectively. The results of the APV-Ab ELISA were classified as positive if the per cent inhibition (PI) exceeded 40%, as borderline if the PI was lower than or equal to 40% and higher than 30%, and as negative if the PI was lower than or equal to 30%. Sera classified as borderline were not included as positives. Examination of the antibody titres against O. rhinotracheale was performed using an indirect ELISA (Intervet International) as described by van Empel et al. (1997).

Statistical analyses

A repeated-measures Poisson regression (Glimmix macro, SAS 8.02) was used to compare the log₂ HI titre profiles between colibacillosis-affected and control flocks for the IBV antigens M41, D274, D1466 and 4/91, for NDV, and for the average O. rhinotracheale titre. A repeated-measures logistic regression (Glimmix macro, SAS 8.02) was used to compare the serological profiles between colibacillosis-affected and control flocks for the proportion of seropositive hens for APV and M. synoviae per flock at the moment of sampling. Age at first sampling (continuous), flock type (colibacillosis-affected versus control flock), the interval between the colibacillosis outbreak and a given sampling (repeated measure, continuous variable), and the interval by flock type interaction were entered in the model as fixed effects, and farm of origin as a random effect. The interval between the colibacillosis outbreak and the first sampling varied for the colibacillosis-affected flocks from 0 (acute outbreaks, quickly reported to our team) to 30 weeks (chronic colibacillosis with a continuous but wavelike increased mortality). For the control flocks, this interval was considered to be the same as for the corresponding colibacillosis-affected flock. A Pearson's chi-square test (Statistix 1.0) was used to compare the total number of seroconversions in the control and the colibacillosis group.

Results

HI antibody titres of the pooled serum samples against IBV M41, D274, D1466 and 4/91 ranged between 7.6 and 9.6 log₂ units (Table 1). No significant differences in antibody titres (P_M41=0.5984, P_D274=0.5096, P_D1466=0.5936, P_4/91=0.8273) were observed between the control flocks and the flocks suffering from colibacillosis. In five flocks, the pooled sera showed a titre increase of two log₂ units when tested with IBV D274 (one control flock) and 4/91 (three control flocks and one colibacillosis flock). The corresponding individual sera yielded mean titre increases ranging from 1.3 to 1.6 log₂ units. Of the five cages tested per flock, one to three showed a mean titre increase of two units or more.

Table 1. HI log₂ titres against IBV and NDV of pooled sera collected with a 6-week interval from 20 flocks of layer hens with colibacillosis and from 20 control flocks

Antigen	Control flocks		Flocks with colibacillosis
	First sampling (minimum to maximum)	After 6 weeks (minimum to maximum)	First sampling (minimum to maximum)	after 6 weeks (minimum to maximum)
IB M41 (8 HA units)	8.4 (6 to 10)	8.6 (7 to 10)	8.6 (6 to 10)	8.5 (7 to 10)
IB D274 (8 HA units)	9.2 (7 to 11)	9.6 (8 to 11)	9.3 (8 to 11)	9.4 (8 to 11)
IB D1466 (8 HA units)	7.8 (6 to 9)	8.0 (6 to 9)	8.4 (6 to 10)	8.1 (6 to 9)
IB 4/91 (4 HA units)	7.6 (6 to 9)	8.2 (7 to 9)	7.9 (6 to 9)	8.2 (7 to 9)
NDV (8 HA units La Sota)	8.6 (7 to 11)	8.5 (7 to 10)	8.5 (6 to 11)	8.8 (6 to 11)

Seroconversion occurred in only one flock, in the control group. In this flock the antibody titre of the pooled sera against IBV D274 rose from 7 to 11 in the successive samplings and was not accompanied by a decrease in egg production or quality.

HI antibody titres of the pooled sera against NDV ranged between 8.5 and 8.8 log₂ units (Table 1). No significant differences in antibody titres (P=0.3019) were observed between control flocks and colibacillosis flocks, and no seroconversion was noted in any of the flocks.

At all samplings in all 40 flocks at least 75% of the birds tested positive in the ELISA for APV antibodies (Table 2). No significant differences in percentage reactors were measured between the control flocks and those suffering from colibacillosis (P=0.5935). In one control flock, an increase of 75% to 100% reactors was seen. In this flock, there was also an increase in the PI of the sera and it was therefore considered to have seroconverted. This observation was not accompanied by a decrease in egg production or quality.

Table 2. Serology for the detection of antibodies or antibody titres against APV, M. gallisepticum, M. synoviae and O. rhinotracheale in sera collected at 2-week intervals from 20 flocks of layer hens with colibacillosis and from 20 control flocks

Pathogen	Test	Flocks	First sampling (minimum to maximum)	Second sampling (minimum to maximum)	Third sampling (minimum to maximum)	Fourth sampling (minimum to maximum)
APV	ELISA per cent reactors	Control	98.0 (75.0 to 100.0)	98.0 (80.0 to 100.0)	98.8 (85.0 to 100.0)	98.5 (85.0 to 100.0)
		Colibacillosis	95.5 (85.0 to 100.0)	95.8 (85.0 to 100.0)	94.8 (80.0 to 100.0)	94.5 (80.0 to 100.0)
M. synoviae^a	ELISA per cent reactors	Control	96.0 (40.0 to 100.0)	95.8 (45.0 to 100.0)	98.5 (80.0 to 100.0)	97.8 (80.0 to 100.0)
		Colibacillosis	86.3 (0.0 to 100.0)	88.1 (5.0 to 100.0)	92.1 (5.0 to 100.0)	93.9 (10.0 to 100.0)
O. rhinotracheale	ELISA titres	Control	9.1 (8.0 to 10.6)	8.8 (7.6 to 10.5)	9.1 (7.4 to 11.4)	9.0 (8.0 to 11.3)
		Colibacillosis	8.7 (7.3 to 9.8)	8.8 (6.9 to 10.6)	8.7 (7.2 to 10.1)	8.7 (7.4 to 9.8)
^a Slide agglutination and ELISA reactions for M. gallisepticum were all negative.

All sera from all hens sampled in the flocks were negative for M. gallisepticum antibodies (Table 2) while all flocks were positive for M. synoviae, with percentages ranging from 85 to 100%, except for one control flock and three colibacillosis-affected flocks. The control flock and two of the colibacillosis flocks showed an increase in percent reactors ranging from 40% to 70% during the follow-up period. Of the two colibacillosis flocks, one was sampled during a first outbreak, and the other during a repeated outbreak (Table 3). In none of these flocks was there a decrease in egg production or egg quality. No significant differences in percentage reactors (P=0.6251) were measured between the control flocks and the flocks suffering from colibacillosis.

Table 3. Time of first sampling, outbreak characteristics and serological results of the colibacillosis-affected flocks

Flock code	Age at first sampling (weeks)	Interval from outbreak to first sampling (weeks)	Number of weeks with increased mortality	Flock sampled at first (1) or repeated (R) outbreak	Increase of titre or number of reactors against
1	21	0^a	3	1	–^b
2	22	2	4	1	M. synoviae
3	22	1	3	1	–
4	25	2	7	1	–
5	26	1	7	1	–
6	29	2	6	1	–
7	33	2	10	1	–
8	36	2	8	R	–
9	36	1	6	1	–
10	36	2	3	R	–
11	36	2	8	R	–
12	38	2	10	1	–
13	39	0	4	R	M. synoviae
14	48	2	3	R	–
15	53	0	4	R	O. rhinotracheale
16	53	2	3	R	–
17	54	1	4	R	–
18	59	2	4	R	–
19	61	2	10	R	–
20	87	2	3	R	–
^a Time interval of less than 1 week.
^b None.

Average flock antibody titres against O. rhinotracheale ranged from 6.9 to 11.3 log₂ units, with a mean of 9.0 and 8.7 in the control flocks and the colibacillosis flocks, respectively (Table 2). Titres of individual birds were as high as 15 log₂ units. No significant differences in titres (P=0.5501) were measured between the control flocks and the flocks suffering from colibacillosis. Seroconversion occurred in one flock of each group, with mean antibody titres in the control flock rising from 8.16 to 11.39 and in the colibacillosis flock from 7.99 to 10.13. The colibacillosis flock was sampled during a repeated outbreak (Table 3). Neither flock showed a decrease in egg production or quality.

When looking at the total number of seroconversions against any of the agents (i.e. four in the control group and three in the colibacillosis group), no significant difference (P=0.6773) was found between the two flock types.

Discussion

The present study examined the presence of antibodies or antibody titres and eventual seroconversion for IBV, NDV, APV, O. rhinotracheale, M. gallisepticum and M. synoviae in 20 layer flocks suffering from colibacillosis-associated mortality and in 20 control flocks of the same age. We hypothesized that if an infection with one of the aforementioned pathogens in the layer flocks took place just before the colibacillosis outbreaks, it would be possible to detect seroconversion in the weeks following the outbreak. The study protocol therefore involved four blood samplings with intervals of 2 weeks. Statistical analysis related the results to clinical and technical data of the flocks. This allowed us to obtain epidemiological data on the prevalence and significance of these respiratory pathogens and their role in the pathogenesis of outbreaks of E. coli mortality in flocks of layer hens. The sampling of 20 hens per flock allowed one only to define within-flock prevalences with a relatively wide 95% confidence interval (CI), using the exact binomial method as described by Clopper & Pearson (1934). If, for example, 10% of the sampled hens were found to react positively, the 95% CI would range from 1.2 to 31.7%. In the case of 50% and 90% reactors, CIs would range from 27.2 to 72.8%, and from 68.3 to 98.8%, respectively. Nevertheless, because a total of 40 flocks was sampled the results were of value, and no significant differences were observed between control and colibacillosis flocks for any of the pathogens under consideration.

In all flocks, antibody titres against NDV and IBV were high as a result of intensive vaccination, possibly boosted by field challenge. However, antibody titres were at levels that can be expected after vaccination. Extremely high titres, indicative of infection with field strains, were not detected, nor were there any IBV-related clinical problems such as a decrease in egg production or quality. Significant differences between the control flocks and the flocks suffering from colibacillosis were not observed. Seroconversion against NDV was not observed at all and only one flock of the control group seroconverted to IBV serotype D274. This may indicate that under the circumstances of this study, NDV and IBV did not predispose flocks of layer hens to outbreaks of colibacillosis.

A high proportion of APV reactors was found in all the sampled flocks. In view of the low frequency of APV vaccination, we conclude that APV infections are common in the layer flocks under study. The APV Svanova ELISA is a qualitative test, but an increase in the PI of the test might indicate a renewed contact of the layers with APV and replication of the virus in the birds.

Antibody titres against O. rhinotracheale were generally high in the colibacillosis flocks as well as the control flocks. Since none of these flocks had been vaccinated against O. rhinotracheale, a high prevalence of O. rhinotracheale infections is indicated in these layer flocks. As far as we know, these are the first epidemiological data on O. rhinotracheale infections in layers in Europe. Sprenger et al. (2000) described mortality and drops in egg production due to O. rhinotracheale infections in flocks of layer hens in North America. During the present study, only one colibacillosis flock seroconverted for O. rhinotracheale. Since this flock was suffering from colibacillosis by the time of sampling, it was not possible to determine whether O. rhinotracheale infection led to increased mortality. The control flock that showed seroconversion had no increased mortality. Drops in egg production or egg quality were not observed. For the remaining seropositive flocks, O. rhinotracheale infections probably occurred at an earlier age, for which serological and technical data were not available.

All flocks were negative for antibodies against M. gallisepticum. Indeed, it was known (M. Verlinden, Animal Health Care Flanders, personal communication) that the prevalence of M. gallisepticum infections in Belgian layers is very low. However, all flocks reacted positively for M. synoviae antibodies, indicating a high prevalence of M. synoviae infection among Belgian layers. The IDEXX MS ELISA is a qualitative test, and since M. synoviae causes a chronic recurrent infection, the number of reactors may increase or decrease. However, no significant differences were found between the control and the colibacillosis flocks, indicating that, under the circumstances of this study, M. synoviae may not have acted as a primary pathogen triggering colibacillosis. The significance of M. synoviae infections in layers has not been well established. Landman & Feberwee (2001) demonstrated a potential role of M. synoviae in the pathogenesis of amyloid arthropathy in layers. In the present study during the time of sampling three flocks seroconverted, one of which belonged to the control group of flocks, and the other two were flocks suffering from colibacillosis. None of the birds of these three flocks, submitted for necropsy, showed lesions of amyloid arthropathy (Vandekerchove et al., in press). No decreased egg production or egg quality was reported for these flocks.

Of the 20 layer flocks suffering from colibacillosis, only three seroconverted against other respiratory agents, notably M. synoviae and O. rhinotracheale. Antibody titres and/or the proportion of birds having antibodies against IBV, NDV, APV, M. synoviae and O. rhinotracheale were similar to the control group of flocks. Thus, even though they may act as a priming agent in certain cases, none of the agents examined were shown to consistently act as a trigger for colibacillosis. This may indicate a primary nature of E. coli infections in flocks of laying hens. In a previous study (Vandekerchove et al., in press), it was demonstrated that E. coli strains involved in typical outbreaks of colibacillosis with high mortality belonged almost exclusively to the serotype O78, while this type was very rarely found in flocks without colibacillosis-associated mortality.

Acknowledgements

The authors wish to thank the farm managers, breeding companies, feed firms, field practicioners and the staff of Flanders Animal Health Care for their valuable help in making this study possible. Also many thanks to D. Vandergheynst, S. Tistaert and M. Decaesstecker for their skilled technical support, to J. de Wit for his helpful advice, and to P. van Empel for organizing the O. rhinotracheale analyses.