You are on page 1of 7

2001 Poultry Science Association, Inc.


Virginia-Maryland Regional College of Veterinary Medicine,
University of Maryland, College Park, MD 20742
Department of Biological Resources Engineering,
University of Maryland, College Park, MD 20742
Department of Cell Biology and Molecular Genetics,
University of Maryland, College Park, MD 20742
Phone: 301-405-5452
FAX: 301-314-9489

Primary Audience: Flock Supervisors, Veterinarians, Extension Agents, Food Scientists,

Quality Assurance Personnel, Epidemiologists.

Broiler farm management should include properly maintained watering devices and appropriate
house ventilation practices for an optimal preharvest production environment. An effective strategy
to prevent dripping water and spillage and to ensure a modest but continuous and uniform
flow of air directly over the litter surface appears to create an unfavorable environment for
enteric bacteria.
Qualitative and quantitative findings have revealed that Salmonella contamination loci are not
equally distributed in the broiler litter. Higher Salmonella and Escherichia coli counts were detected
in litter samples with water activity (aw) greater than 0.90 and percentage moisture content (MC)
greater than 35%. At reduced aw and MC levels, the numbers of viable Salmonella cells were low,
further reflecting the importance of preventing excessively damp areas (e.g., “cake”) in broiler
litter. Additionally, litter surface airflow measurements suggested that those areas exposed to
modest air velocities (100 to 150 ft/min or approximately 1.5 mph) had drier litter and reduced
E. coli populations.

Key words: Airflow, litter management, litter moisture, risk reduction, Salmonella, water activity
2001 J. Appl. Poult. Res. 10:245–251

Presently at Research Triangle Park, NC 27709.
To whom correspondence should be addressed.
246 JAPR: Research Report

entire house, was sampled using four to six DS

DESCRIPTION OF PROBLEM per chamber to obtain an overall perspective of
High water parameters in the litter can create the Salmonella status of the house. Swabs were
favorable conditions for the multiplication of placed in 1,000-mL Whirl-Pak bags and then
enteric pathogens such as Salmonella and E. placed inside an insulated, refrigerated box. The
coli. Indeed, previous drag swab (DS) studies swabs were processed, and broth suspensions
[1, 2] have shown that broiler litter surfaces were cultured qualitatively as described by
with water activity (aw) greater than 0.90 were Opara et al. [1]. Next, 96 litter samples were
associated with increased Salmonella recovery collected as described by Mallinson et al. [4].
rates. Furthermore, the risk of Salmonella con- Each sample comprised 5 to 10 pooled collec-
tamination on processed broiler carcasses is re- tions from wet and dry areas (five samples from
duced when carcasses originate from farms with areas under nipple drinkers and five from areas
low detectable levels of Salmonella [3]. Com- 5 ft away). These areas were selected because
bined, these observations suggest that lower aw they had comparable moisture levels between
levels at the litter surface likely contribute to houses on the same farm and between farms.
lower Salmonella loads on carcasses [3]. There- Furthermore, pooled samples were used to in-
fore, the transmission of Salmonella from the crease the probability of detecting Salmonella
farm to the processing plant and potentially to from the litter. These pooled litter samples were
marketed carcasses may be diminished and con- collected and placed into 36-oz Zip-Lock bags,
trolled by implementing management strategies using separate disposable surgical gloves for
that reduce bacterial loads in the production en- each pool. Samples were collected from identical
vironment. Identification of Salmonella and E. sites for physical (aw and MC) and bacterial
coli multiplication “hot spots” in the litter should (Salmonella and E. coli) analyses.
enhance poultry health and quality control.
This study, based on the quantitative Salmo-
nella and E. coli data obtained from litter sam- Airflow patterns over each sampled area
ples with various levels of aw and moisture con- were measured using a Digital Hygro thermome-
tent (MC), identified regions in litter that consti- ter, anemometer, data-logging instrument ac-
tute multiplication “hot spots” (i.e., litter with cording to the manufacturer’s instructions [5].
aw from 0.90 to 0.95 or MC from 35 to 50%). The instrument’s anemometer, located at the end
The measurement of airflow velocities over litter of a one-meter handle bar, was held approxi-
surfaces further revealed that moderately in- mately 1 in. from the litter base at the center of
creased, rates of airflow directly over the litter the area from which a particular litter pool was
surface generally contribute to drier litter and, collected. Each air velocity reading represented
consequently, to decreased viable Salmonella the average flow of air at a particular location
and E. coli populations. during a 15-s interval.
The aw and MC were measured as previously
Brood chambers of 13 commercial broiler described [6]. Litter samples were processed ac-
houses from four different farms located on the cording to Mallinson et al. [4]. Briefly, litter
Delmarva Peninsula were visited on a single or samples and DS were immediately transported
on multiple occasions during the growout pe- back to the laboratory in an insulated container
riod, except for one farm, which was visited for processing. Thirty grams from each litter
during a subsequent growout period. On all sample was suspended in 300-mL of buffered
farms, except for one, sampling was restricted peptone water (BPW) contained in 500-mL
to the brood chambers, because preliminary data polypropylene wide-mouth bottles [7]. Samples
(not included) indicated that they had the highest were vigorously shaken for 2 min and lightly
prevalence of Salmonella-positive DS. At each filtered. Three 80-mL aliquots were divided
visit the brood chamber and, in some cases the among three 100-mL Snap-Cap jars [8], and two

10-mL aliquots of the filtrate were poured into nella spp. were not uniformly distributed
two 15-mL centrifuge tubes. The jars and tubes throughout the litter surface. In this study, Sal-
were immediately flash-frozen at −70°C. monella spp. were detected in only 17.7% of
all collected litter samples (Table 1). The DS
SALMONELLA QUANTIFICATION technique proved to be more sensitive than cul-
Litter samples were prescreened for further ture of litter samples for establishing the Salmo-
Salmonella quantification as previously de- nella status of a broiler house with 2.4× greater
scribed [4]. Briefly, Salmonella screening was detection (42.5%) in commercial houses (Table
performed by qualitatively testing for this organ- 1). Because the DS method involves sampling
ism using two of the three 80-mL litter suspen- from a large surface area in the house, it is more
sions for each sample. The purpose of prescreen- likely to be in contact with the contaminated
ing was to preserve materials and to avoid per- areas.
forming the laborious five-tube most probable Although the DS method helps establish,
number (MPN) procedure on samples that were semi-quantitatively, the general prevalence of
essentially Salmonella-negative. The last 80-mL Salmonella contamination in the house litter, it
suspension of those samples positive for Salmo- does not adequately examine the direct contribu-
nella at prescreening were quantified for Salmo- tion of physico-chemical litter parameters on
nella by using the five-tube MPN procedure [9, Salmonella and E. coli populations. By collect-
10]. After primary and selective enrichment in ing data on specific conditions and correspond-
BPW and tetrathionate-Hajna [11] broth, respec- ing bacterial populations of litter samples, the
tively, all MPN tubes were plated on Miller- individual or cumulative effects of certain pa-
Mallinson (MM) [12] agar and incubated at rameters (i.e., aw and airflow) on bacterial
35°C. Plates were read after 24 and 48 h of incu- growth can be identified. This approach can help
bation. determine the optimal environmental conditions
for the production of healthier and safer poultry.
E. COLI QUANTIFICATION Previous studies [1, 6] have established that
E. coli quantification was performed by a relationship exists between litter regions with
thawing one 10-mL aliquot suspension of litter, high aw levels (>0.90) and the number of Salmo-
which was serially diluted and spread-plated (0.1 nella-positive DS. However, the data presented
mL) on eosin-methylene-blue agar (EMB) [11]. here further suggested that high litter MC also
The plates were incubated at 35°C for 24 hr contributed favorably to the presence of Salmo-
and typical E. coli colonies were counted. The nella. Therefore, understanding the relationship
smaller (10-ml) litter suspension aliquot was between these two water parameters becomes
used for E. coli quantification because of the important because both may serve as strong indi-
higher load of this organism in the litter. cators of high-risk conditions that support Sal-
monella contamination. We observed a curvilin-
STATISTICAL ANALYSIS ear relationship between aw and MC levels in
We determined the significant difference of broiler litter (Figure 1). As MC levels decreased
the means of aw, MC levels, and E. coli popula- from approximately 60 to 35%, aw decreased
tions for litter samples collected from areas ex-
posed to varying airflow levels by using the TABLE 1. Comparisons between drag swab and litter
directional Student’s t test [13]. sampling for the detection of Salmonella on seven
broiler farmsA


As reported by Hayes et al. [14], the DS
Drag swab (n = 40) 42.5
method was more reliable for qualitatively de- Litter samplesB (n = 96) 17.7
termining the Salmonella status of a house than
culturing litter samples. When compared to the The only houses that were included in the table were those
where Salmonella was detected by either or both methods.
overall high percentage of negative litter culture B
There was only one instance when a house was positive
results, these observations suggested that Salmo- by litter sampling but negative by the drag swab method.
248 JAPR: Research Report

average recovery was 2.7. This finding provided

an explanation for the sporadic positive cultures
obtained in previous studies by Opara et al. and
Carr et al. [1, 2], who reported significantly
lower DS Salmonella recovery rates from broiler
farms having average aw less than 0.85 in the
In a qualitative study encompassing litter
samples and DS taken from a total of 86 com-
FIGURE 1. Relationship between water activity (aw) mercial farms on the Delmarva Peninsula, some
and percentage moisture content (%MC) in broiler litter Salmonella-positive samples with low aw levels
samples (n = 96).
were found but were not quantified [14]. Of a
total of 29 Salmonella-positive litter cultures,
gradually almost linearly. Further, as MC levels 34.5% had aw values <0.85. Detection of low
decreased to <35%, aw values sharply decreased. levels of viable Salmonella cells may reflect the
This observation was significant because it re- presence of organisms poised to multiply, should
vealed that the MC value might also be important a more favorable condition occur (e.g., addition
when attempting to describe the conditions of of moisture to the litter bed or ventilation rates
litter that support bacterial growth. MC reflects or distribution patterns incapable of adequately
the effectiveness of ventilation in drying litter removing moisture from the litter). This finding,
[15] and in Salmonella contamination rates [16]. showing a direct Salmonella quantitative rela-
The MC level at a particular litter location can tionship to aw and MC, reinforced our previous
possibly maintain a fairly constant aw by releas- hypothesis by providing further evidence that
ing water molecules at the expense of total MC.
However, a point is reached at which unbound
water on a litter matrix is diminished, thus caus-
ing a sharp drop in aw. A similar trend between
aw and MC has been reported for fecal drop-
pings [17].
Figure 2 represents Salmonella-positive lit-
ter samples and respective Salmonella popula-
tions grouped into ranges according to aw and
MC values. In Figure 2a, results suggest that
litter samples with aw between 0.90 to 0.95 may
represent an optimal situation for the growth of
Salmonella (average of 44.7 viable Salmonella
colony-forming units recovered from 10 g of
litter collected). Furthermore, the data revealed
0.85 to 0.89 aw to be the second best condition
for growth of Salmonella (average count of 20.7
cfu/10 g). In this series, one sample, which had
an aw of 0.893, had an unusually high Salmonella
count of 540 cfu/10 g of litter. This unusually
high count was largely responsible for the high
final average calculated for this aw range.
Not surprisingly, Salmonella was detected
in low numbers in litter samples with aw <0.85.
Quantification results revealed that the number
of viable cells was extremely low. The average
FIGURE 2. Quantification of Salmonella spp. from litter
recovery of Salmonella from 10 g of litter with samples grouped according to A) water activity (aw) and
aw from 0.84 to 0.80 was 0.13; with aw <0.79 B) percentage moisture content (%MC) in broiler litter.

Analysis of the culture results in relation

to MC provided information analogous to that
observed with aw data (i.e., MC levels between
35 and 50% were more frequently associated
with higher Salmonella counts) (Figure 2b).
Therefore, 35 to 50% MC in the litter appears
to constitute a high-risk situation for the pres-
ence of Salmonella in large numbers (average
of 250 viable Salmonella cells/10 g of litter).
Significantly, culture quantification results re-
vealed that Salmonella populations might be 50-
to 500-fold higher if the litter bed aw is between
0.90 to 0.95 or MC levels are between 35 to
50% (Figure 2).
Parallel to the Salmonella quantification re-
sults, quantification of E. coli populations re-
vealed a similar trend (Figure 3). Litter with aw
between 1.0 to 0.95 and 0.949 to 0.90 had the
highest average E. coli populations per 10 g of
litter (3.74 × 106 cfu/10 g litter and 3.58 × 106
cfu/10 g litter, respectively). Compared to the
FIGURE 3. Quantification of Escherichia coli from litter average viable E. coli population for these two
samples grouped according to A) water activity (aw) and
B) percentage moisture content (%MC) in broiler litter. aw intervals, reductions by 60.9 and 68.2% were
observed for samples with aw levels ranging
from 0.90 to 0.85 and below 0.849, respectively.
Salmonella could possibly be surviving and per- Based on litter percentage of MC, lower E.
sisting in the litter bed in a viable but non- coli populations were also obtained for samples
culturable phase [18, 19]. Regardless, growth with MC <35% as compared >35% (1.0 × 106
conditions in the litter, such as elevated water cfu/10 g of litter vs. 6.67 × 106 cfu/10 g of litter).
parameters, might promote greater survival and This difference in average count represents an
notable multiplication of the organism. 85.0% reduction in the number of viable E. coli
Interestingly, litter samples with water pa- cells for samples with MC <35%. This compara-
rameters in the highest aw range (0.95 to 1.00) tively lesser effect from that noted for Salmo-
were not associated with elevated Salmonella nella could be due to the constant reintroduction
populations. Possibly such conditions are less of E. coli into the litter bed from continuous
than ideal for the growth of Salmonella, perhaps fecal shedding by birds.
due to the creation of a hypotonic environment Airflow results indicated that areas exposed
or to the dilution of essential nutrients. to higher ventilation rates (above 60 ft/min) were

TABLE 2. Airflow, water activity (aw), percentage moisture content (%MC), and Escherichia coli populations for
litter surfaces of six commercial broiler flocks

FPMA (n = 92B) (FPM) %MCC awD 10.0 g LITTERE

<60 (58) 27.19 41.17 0.909 5.4 × 106

≥60 (34) 155.2 28.59 0.849 8.2 × 105
Feet per minute.
Average of 92 samples. Four litter samples were not used because airflow measurements were not taken.
MC = percentage moisture content. Means within the column differ (P < 0.05).
aw = water activity. Means within the column differ (P < 0.05).
E. coli count obtained using serial dilutions and spread-plating on eosin-methylene-blue agar. Means within the column
differ (P < 0.10).
250 JAPR: Research Report

significantly associated with lower MC, aw, and the ft/min of airflow (data not shown). Therefore,
E. coli population [13]. Because there were a in addition to already reported advantages of
limited number of Salmonella-positive litter tunnel ventilation, such as a reduction of heat-
samples, the direct impact of ventilation on the stress related mortality and improved feed con-
reduction of Salmonella populations was not de- version [20], tunnel ventilation could also be
finitively established. Due to a similarity in re- advantageous as an effective means of removing
sponse to aw and MC changes, the results ob- excess moisture from litter, thereby helping to
tained for E. coli can be extrapolated to Salmo- suppress Salmonella and E. coli levels and risks.
nella (Figure 2). Although ventilation practices For growers attempting to avoid the higher in-
varied considerably between farms (number, stallation and operational costs associated with
placement, static pressure, running time of fans, tunnel ventilation [21], but concerned about ob-
and curtain setting), a correlation seemed to exist taining a more homogenous airflow coverage
between the rate of air movement over a specific over the litter bed in naturally ventilated houses,
litter location and its water parameters (Table ceiling fans or perhaps even baffles may provide
2). Locations where air movement was approxi- an attractive alternative to side-wall propeller
mately zero tended to have higher water parame- fans [21].
ters, which might have increased the likelihood This study provided additional evidence that
of pronounced Salmonella multiplication in such aw and MC in the litter and air velocity over
areas, should the organism have been in- the litter represent potentially important house
troduced. parameters that need to be monitored to ensure
It was also noted that, in contrast to houses a safe and healthy environment for broiler pro-
that were ventilated by wind and propeller fans, duction. By knowing the levels of contamina-
tunnel ventilation houses (two observations) had tion, one can better determine the degree of con-
a higher (131 to 198 ft/min) and unvarying flow tamination and, thus, take appropriate measures.
of air over the entire litter surface in the brooder Such measures could involve improving the uni-
chamber. This elevated and continuous flow of formity of moderate airflow rates of 100 to 150
air might have contributed to the low detection ft/min (approximately 1.5 mph) and repairing
of Salmonella from litter samples (1 positive of roof leaks and leaking watering devices. While
26 collected). To achieve airflow of this magni- avoiding establishment of undesirable dry dusty
tude (150 to 200 ft/min) the grower had four conditions, the corrective management practices
tunnel fans running concurrently, generating an suggested here might likely produce those litter
in-house static pressure between 0.04 to 0.05. parameters associated with reduced risks of E.
Static pressure appears to be directly related to coli and Salmonella contamination.


1. Careful attention must be given to aw and MC levels in broiler litter. According to recent data,
aw <0.85 and MC <35% are recommended.
2. Any water leakage must be repaired immediately so as to prevent addition of moisture to the
litter base.
3. Ventilation should be managed to maintain a litter drying environment, which can be achieved
by regularly maintaining a modest and uniformly distributed ventilation rate (100 to 150 ft/
min) over the litter surface.


1. Opara, O.O., L.E. Carr, E. Russek-Cohen, C.R. Tate, 2. Carr, L.E., E.T. Mallinson, C.R. Tate, R.G. Miller, E.
E.T. Mallinson, R.G. Miller, L.E. Stewart, R.W. Johnson, and Russek-Cohen, L.E. Stewart, O.O. Opara, and S.W. Joseph, 1995.
S.W. Joseph, 1992. Correlation of water activity and other environ- Prevalence of Salmonella in broiler flocks: Effect of litter water
mental conditions with repeated detection of Salmonella contamina- activity, house construction, and watering devices. Avian Dis.
tion on poultry farms. Avian Dis. 36:664–671. 39:39–44.

3. Mallinson, E.T., L.E. Carr, G.W. Malone, C.A. Wabeck, water activity and moisture content to the population distribution of
D.H. Palmer, E.B. Pusey, E. Russek-Cohen, and S.W. Joseph, Salmonella spp. in commercial poultry houses. Poult. Sci.
1995. Lower water activity in broiler litter and the reduction of 79:1557–1561.
Salmonella on farms and processed carcasses. Pages 1–12 in: Bull.
348. Maryland Cooperative Extension Service, College Park, MD. 15. W.D. Weaver, Jr., and R. Meijerhof, 1991. The effect
of different levels of relative humidity and air movement on litter
4. Mallinson, E.T., C. E. de Rezende, N.L. Tablante, L. E. conditions, ammonia levels, growth and carcass quality for broiler
Carr, and S.W. Joseph, 2000. A management technique to identify chickens. Poult. Sci. 70:746–755.
prime locations of Salmonella contamination on broiler and layer
farms. J. Appl. Poult. Res. 9:364–370. 16. Jones, F., R.C. Axtell, F.R. Tarver, D.V. Rives, S.E.
Scheideler, and M.J. Wineland, 1991. Environmental factors con-
5. Pacer Industries, Inc., Model No. HTA-4200, Chippewa tributing to Salmonella colonization in chickens. Pages 3–20 in:
Falls, WI 54729. Colonization Control of Human Bacterial Enteropathogens in Poul-
6. Carr, L.E., E.T. Mallinson, L.E. Stewart, and S.W. Jo- try. L.C. Blankenship, ed. Academic Pres, Inc., San Diego, CA.
seph, 1994. The influence of water activity on Salmonella presence
17. Himathongkham, S., S. Nuanualsuwan, and H. Riemann,
in broiler operations. Appl. Eng. Agric. 10:403–405.
1999. Survival of Salmonella enteritidis and Salmonella typhimurium
7. Nalgene, Rochester, NY 14830. in chicken manure at different levels of water activity. FEMS Micro-
8. Snap-Seal Container, No. 1730 series, Science Product Div., biol. Lett. 172:159–163.
Corning, Inc., Corning, NY 14830. 18. Chmielewski, R.A.N., and J.F. Frank, 1995. Formation of
9. Mallinson, E.T., R. G. Miller, R.M. Dyer, D.J. Holder, viable but nonculturable Salmonella during starvation in chemically
C.M. Lamichhane, M.L. Giddens, E. Russek-Cohen, J. deGraft- defined solutions. Lett. Appl. Microbiol. 20:380–384.
Hanson, and S.W. Joseph, 1995. Quantification of Salmonella from 19. Roszak, D.B., D.J. Grimes, and R.R. Colwell, 1984. Viable
carcass rinse and environmental specimens using a membrane filter but nonrecoverable stage of Salmonella enteritidis in aquatic systems.
transfer method. Proc. Salmonella Symp. 99th Annu. Mtg. U.S.
Can. J. Microbiol. 30:334–338.
Anim. Health Assoc. and 38th Annu. Conf. Am. Assoc. Vet. Lab.
Diagnosticians, Reno, NV. 20. Lacy, M.P., and M. Czarick, 1992. Tunnel-ventilated
10. Food and Drug Administration, 1992. Bacteriological Ana- broiler houses. Broiler performance and operating costs. J. Appl.
lytical Manual. 7th ed. FDA, Washington, DC. Poult. Res. 1:104–109.

11. Becton-Dickinson, Cockeysville, MD 21030. 21. Bottcher, R.W., J.R. Magura, J. S. Young, and G. R.
Boughman, 1995. Effects of tilt angles on airflow for poultry house
12. Mallinson, E.T., R.G. Miller, C.E. de Rezende, K.E. mixing fans. Appl. Eng. Agric. 14:87–96.
Ferris, J. deGraft-Hanson, and S.W. Joseph, 2000. Improved plat-
ing media for the detection of Salmonella species with typical and
atypical hydrogen sulfide production. J. Vet. Diag. Invest. 12:83–87. ACKNOWLEDGMENTS
13. Samuels, M.L., and J.A. Witmer, 1999. Comparison of We thank Russell Miller for his valuable technical assistance and
two independent samples. Pages 234–311 in: Statistics for the Life Evandro Missagia for hardware and software support. This research
Sciences. 2nd ed. Prentice Hall, Upper Saddle River, NJ. was supported by a grant from the U.S. Poultry and Egg Association.
14. Hayes, J.R., L.E. Carr, E.T. Mallinson, L.W. Douglass, This research contributed to Regional Poultry Research Project
and S.W. Joseph, 2000. Characterization of the contribution of NE-127.