Top Navigation Bar

ASAE Conference Proceeding

This is not a peer-reviewed article.

Effectiveness of Vegetative Filter Strips (VFS) for Controlling Pathogen Loads and Antibiotic Resistance in Dairy Wastewater

E. Frantz, K. Griswold, G. Apgar, B. Jacobson, and J. Haddock

Pp. 067-073 in the Animal, Agricultural and Food Processing Wastes, Proceedings of the Ninth International Symposium, 11-14 October 2003 (Raleigh, North Carolina, USA), ed. Robert Burns. ,11 October 2003 . ASAE Pub #701P1203

Abstract

A vegetative filter strip (VFS) was assessed for controlling pathogen loads and antibiotic resistance (AR) in dairy wastewater. The VFS was divided into two equal sections, and wastewater was applied to alternating sections in a crossover design every 6 months. Dairy wastewater and surface water surrounding the VFS and a negative control (NC) site were collected every two weeks and after rain events (>1 in. per 24 h) during the study. Fecal coliform (FC) and fecal streptococci (FS) were measured as pathogen load indicators, and AR of the FS was tested with the antibiotics: tetracycline, oxytetracycline, neomycin, streptomycin, erythromycin, ciprofloxacin, and vancomycin. Pathogen loads (CFU log 10 per 100 mL) and AR data were analyzed using SAS MIXED procedures, and the model included site, season, rain event, effluent application, and the interaction of site by season. Data on AR of FS from animals was analyzed using the SAS MIXED procedures with the model including species. Site and season affected FC and FS loads ( P < 0.05). Fecal coliform loads decreased from wastewater (10 6 ) to surface waters (10 1 -10 5 ). Fecal streptococci loads decreased from wastewater (10 8 ) to surface waters (10 2 -10 6 ). Antibiotic resistance was affected by sampling site and season ( P < 0.05). Streptomycin and neomycin exhibited the greatest levels of AR while vancomycin exhibited the least. Movement of AR to streptomycin, tetracycline and oxytetracycline may have been affected by animals that habitat the VFS and not the dairy cattle. VFS systems can effectively reduce pathogen loads in dairy wastewater.

KEYWORDS. pathogen loads, antibiotic resistance, vegetative filter strip, dairy, wastewater.

Introduction

In light of the current concerns over contamination of water supplies by waste from animal feeding operations (AFOs), it would be prudent to determine the effectiveness of existing technologies for protecting water resources. On dairy farms, wastewater from feedlot runoff and cleaning of milking facilities is a concentrated source of bacterial pathogens (Muchmore et al.,1976). Vegetative filter strip (VFS) systems have been employed as wastewater treatment facilities on small dairy farms (e.g. <300 cows) in Illinois for over 25 years (Muchmore et al.,1976). However, beyond the first few years of operation, there is little if any data on the effectiveness of these systems (Muchmore et al., 1976).

Pathogenic strains of Salmonella, Escherichia coli and Streptococci are present in the gastrointestinal tract of warm-blooded animals and may cause diseases in humans (Wiggins, 1996; Toranzos and McFeters, 1997). There is concern that the use of antibiotics in animal feed and for treatment of diseased animals is selecting for antibiotic resistant pathogens, which will reduce the effectiveness of certain antibiotics (FDA, 1998). Thus, controlling and preventing the movement of bacterial pathogens and the antibiotic resistance genes that they may carry from AFOs to the environment takes on great importance depending on the distance to nearby water systems. Therefore, with the use of a 25-year old VFS system that has been in continual use on an operating dairy farm, the objectives of the current study were to determine: 1) The effect of effluent distribution on the levels of pathogen indicator organisms in water surrounding the VFS system over time, and 2) The antibiotic resistance patterns of fecal streptococci in and around the VFS system.

Materials and Methods

VFS System and Operation

The current study was conducted using the 25-year old VFS system at the Southern Illinois University (SIU) Dairy Center located in Carbondale, IL. The study was conducted from May 15, 2001 to June 15, 2002, and during that time, effluent from the VFS system was distributed within a 61 m x 61 m square centered on the diagonal over a naturally occurring ridge in the field application area (FAA). Centering the square diagonally on the ridge line created two equal area triangles or plots. For 6.5 months, effluent was applied to only the southeast plot while the northwest plot was left dormant (i.e. received no effluent), and then, for the remaining 6.5 months, effluent was applied to the northwest plot while the southeast plot was left dormant. Within each plot, the distribution manifold was moved to a randomly selected location every 2 weeks to provide complete coverage of the plot with effluent and prevent burning of the vegetation by excessive buildup of N.

The vegetation in the FAA consisted primarily of fescue with Reeds Canarygrass along the peripheries. Vegetation was harvested from the FAA three times during the study using a diskbine set to leave 10 cm of growth, and a round-bale baler. The distribution manifold was moved to an adjacent grass lot 2 days prior to and returned 1 day post harvest. Based on prior measurements, under normal conditions without rainfall, the estimated volume of effluent pumped onto the FAA was approximately 1100 to 1500 L day -1 . The maximum amount of effluent pumped onto the FAA day -1 depended on rainfall throughout the time of the study.

The diet of the 50 cow herd at the SIU Dairy Center was a constant ratio of 50:50 forage:concentrates that was formulated to meet the nutrient requirements of a lactating, Holstein, multiparous cow producing 35 kg of 3.8% fat milk. Cows were milked twice day -1 , and waste milk along with the wastewater from cleaning the milking system drained directly to the cistern of the VFS system.

Sampling and Measurements

Water and wastewater samples were collected every two weeks and if possible after every major rain event (>2.54 cm 24 h -1 ). Water samples (1 L) were collected by hand into sterile labeled sealable plastic bottles (Fisher Scientific, St Louis, MO) from 5 collection sites and the manifold (Figure 1). Water samples were also collected from a negative control site located approximately 1.6 km south of the VFS system in a lowland area that drained into the Carbondale reservoir. Samples were stored on ice for transport, and were processed within 6 h of collection.

Fecal samples (100 g) from 10 randomly selected cows were collected seasonally to determine FC and FS counts along with AR patterns of FS. Feces were collected into individual sterile whirl packs (Fisher Scientific, St Louis, MO) by direct collection as the cow defecated. Fecal samples (100 g) from the local wild deer herd that traverse the VFS system were collected into sterile whirl packs from fresh dung piles, and samples of gastro-intestinal tract contents from 10 randomly selected wild pigeons that inhabited the SIU Dairy Center were collected after sacrifice and aseptically transferred to clean plastic weigh boats, weighed, and sub-samples (1 g) were collected into sterile 15 mL Falcon tubes (Fischer Scientific, St Louis, MO). All fecal and gastro-intestinal tract samples were processed within 6 h of collection.

Bacterial Counting and Determination of Antibiotic Resistance

During processing, water samples were sub-sampled into 10 mL, 1 mL, and 0.1 mL aliquots and filtered onto 100mL Nalgene Analytical Test Filter Funnels (Fisher Scientific, St Louis, MO). Wastewater samples were sub-sampled, serially diluted to 10 ¯7 and filtered onto micropore filters, and filters were placed into sterilized Millipore Petri dishes with pads. Filters and Petri dishes were incubated for 24 h at 44.5ºC in M-FC broth for FC and for 48 h at 35ºC in K F Streptococcal broth (BBL, Fisher Scientific, St Louis, MO) for FS . Colonies were counted for each plate, and any plates that contained > 200 colonies were termed “too many to count” (TMTC). Plates containing 20 to 60 colonies were considered countable and used to determine bacterial density. If any plate in a given dilution series did not contain 20 to 60 colonies, then all plates with <200 colonies were combined and divided by the total amount of sample filtered for those plates to determine bacterial density. Bacterial densities were reported as colony forming units (CFUs) per 100 mL of sample.

Ten FS colonies were randomly selected and isolated from each available water, wastewater or fecal sample for determination of AR. Bacterial isolates were spread on plates of tryptic soy agar to produce a bacterial lawn. Two lawns of each isolate were used for testing AR. Antibiotics tested included: ciproflaxacin, vancomycin, erythromycin, neomycin, streptomycin, tetracycline, and oxytetracycline. Antibiotic disks (Fisher Scientific, St Louis, Mo) were placed equidistant apart on the bacterial lawns. One plate contained vancomycin (30 mg), neomycin (30 mg), and tetracycline (30 mg), and the other contained ciprofaxcin (5 mg), erythromycin (15 mg), streptomycin (10 mg), and oxytetracycline (30 mg). Plates were inverted and incubated at 29 °C for 48 h. Antibiotic resistance was measured as the diameter of the inhibition zone around each disk according to the manufacturer’s guidelines and was expressed as susceptible = 0, intermediately resistant = 1, or resistant = 2.

Experimental Design and Statistical Analysis

The study was designed as a crossover split plot nested within a randomized complete block with repeated measures. Data were analyzed using the PROC GLM and PROC MIXED procedures of PC SAS Version 8.0 (SAS Institute, Inc., Cary, NC). For pathogen indicator and AR data collected from water samples, the statistical model included the fixed effects of site, season, rain event, effluent application, and the interaction of site by season. The model as fitted was: Y ijk = µ + S i + SE j + R k + E m + SSE ij + e ijk . Where, Y ijk is the dependent variable, µ is the overall mean, S i is the fixed effect of the i th site (i = 1, …, 6), SE j is the fixed effect of the j th season (j = 1, …, 4), R k is the fixed effect of the k th rain event (k = 1, 2), E m is the fixed effect of the m th effluent application (l = 1, 2), SSE ij is the fixed interaction effect of the i th site by j th season, e ijkl is the residual error. Date, pH and air temperature were used as covariates in the model. Where applicable, single df orthogonal contrasts were included to test: manifold vs. negative control, manifold vs stream sites, negative control vs VFS stream sites, stream site 2 vs stream site 5, manifold vs stream site 5, spring & summer vs fall & winter, and winter & spring vs summer & fall. Significance was determined at P < 0.05.

For AR data collected from animal samples, data were analyzed by antibiotic, and the statistical model included the fixed effects of species. Where applicable, single df orthogonal contrasts were included to test: cows vs deer, cows vs pigeons, pigeons vs deer, cows & pigeons vs deer, and cows & deer vs pigeons. Additionally, AR data for FS from animals and water samples were analyzed using the PROC DISCRIM procedures of PC SAS. PROC DISCRIM classifies data into categories according to patterns within dependent variables.

Results and Discussion

Pathogen Indicator Bacterial Loads

Fecal coliform (FC) and fecal streptococci (FS) concentrations in water samples are used as indicators of potential microbial pollution (Csuros and Csuros, 1999). The concentrations of FC and FS found in the manifold effluent and stream water samples are provided in Tables 1 and 2. The values given are the log 10 transformations of the CFU per 100 mL of sample. The concentrations of FC in the manifold effluent ranged from approximately 25,000 to 7,080,000 CFU/100 mL while FS ranged from approximately 2,040,000 to 135,000,000 CFU/100 mL during the study (Tables 1 and 2). The concentrations of FC in the stream samples ranged from approximately < 10 to 480,000 CFU/100 mL while FS ranged from approximately 120 to 3,300,000 CFU/100 mL during the study (Table 1 and 2). The significant changes (P < 0.0001) in pathogen indicators represent a 100 to 10,000 fold decrease in the number of pathogens in the water samples, which clearly indicated that the VFS system was effectively removing pathogens from the effluent prior to it entering the surrounding stream waters. Further, FC concentrations in the VFS stream sampling sites were not different (P > 0.05) from the control stream site, which is located approximately 1.6 km south of the VFS and not exposed to agricultural runoff. Season significantly affected the concentrations of FC ( P = 0.0002) and FS ( P = 0.0003) in both manifold effluent and stream water samples with concentrations being lower in the winter and spring compared to the summer and fall (Tables 1 and 2).

Antibiotic Resistance Patterns

Antibiotic resistance (AR) of pathogens has and continues to be an increasing concern in terms of public health in the U.S. In order to assess the degree of AR movement from dairy wastewater to the surrounding environment, the resistance of FS to ciprofloxacin, vancomycin, erythromycin, streptomycin, neomycin, tetracycline, and oxytetracycline was determined. All of the antibiotics except for vancomycin exhibited a change in the degree of AR with changing seasons (Table 3). These changes in AR were presumably related to changes in the bacterial populations, and may suggest that AR affects the ability of certain FS to survive when environmental conditions change due to season.

The resistance of FS to the antibiotics, streptomycin, tetracycline, and oxytetracycline, was affected by sampling site (Table 4). In all three cases, AR was greater ( P = 0.01) in the VFS stream sites than in the control stream site. This difference in AR likely reflects the contribution of FS from the dairy wastewater that are resistant to these antibiotics to the VFS stream, however, none of the three antibiotics has been used at the SIU Dairy Center in the past 20 years. Therefore, where did the AR FS originate from? Fecal streptococci from dairy wastewater in the VFS would presumably originate from the cows in the SIU dairy herd; however, other animals can contribute to the FS population. In order to assess the potential contribution of species other than the cows, fecal samples from cows, deer and pigeons that habitat the area were collected, and AR of the FS was determined for the same seven antibiotics. Initially, a statistical classification procedure was used to determine the contribution of each species to the AR patterns of the FS isolated from the VFS system. This analysis was able to accurately classify (96%) of the FS isolates from cows and deer, but could not accurately classify (77%) of the FS isolates from pigeons (Table 5). Therefore, results of classifying unknown FS isolates from the manifold, VFS stream and control stream are questionable. Additional analysis of the AR patterns of the FS isolates from the cows, deer, and pigeons indicated that the resistance of cows and pigeon FS isolates to streptomycin, tetracycline and oxytetracycline was not different (Table 6), which indicated that pigeons may contribute to AR of pathogens in and around the VFS system. The pigeon flock tested migrates from the SIU Dairy Center to the other SIU animal research centers on a daily basis, and the SIU Swine Center, which is 0.8 km from the SIU Dairy Center, still uses oxytetracycline and tetracycline in some feeds. It may be possible that the pigeons consume the antibiotics at the SIU Swine Center, and then, defecate either the antibiotic or AR FS into the ration consumed by the cows at the SIU Dairy Center. As the SIU Dairy Center and Swine Center do not share equipment, personnel or feedstuffs, the pigeon flock is the only common link between the two sites and may perpetuate the movement of AR from dairy wastewater to the streams surrounding the SIU Dairy Center.

Conclusion

The results of this study indicate that older VFS systems can effectively remove pathogen pollutants from dairy wastewater. The VFS system in this study was able to reduce pathogen indicator organism concentrations by 100 to 10,000 fold. The movement of antibiotic resistant pathogens through the VFS systems may be dependent on species other than the cows producing the wastewater since bird populations have been shown to carry pathogens with the same antibiotic resistance patterns. Future research should focus on using biosecurity such as fencing or bird control to reduce the movement of antibiotic resistant pathogens through the VFS system.

REFERENCES

Csuros, M. and C. Csuros. 1999. Microbiological Examination of Water and Wastewater . Boca Raton, FL: Lewis Publishers.

FDA. 1998. Guidance for industry: evaluation of the human health impact of the microbial effects of antimicrobial new animal drugs intended for use in food-producing animals. Fed. Reg. 63(222), Nov. 18.

Muchmore, C. B., E. E. Cook, and M. J. Battaglia. 1976. Land application of agricultural wastewater. In Proceedings from the 31 st Purdue Industrial Waste Conference, 676-683. Ann Arbor Science, MI.

Toranzos, G. A., and G. A. Feters. 1997. Detection of indicator microorganisms in environmental freshwaters and drinking waters. In Manual of Environmental Microbiology . Washington, D.C.: ASM Press.

Wiggins, B. A. 1996. Discriminant analysis of antibiotic resistance patterns of fecal streptococci, a method to differentiate human and animal sources of fecal pollution in natural waters. Appl. Environ. Microbiol. 62:3997-4002.

Appendix

Table 1. Treatment least squares means for the effects of sampling site and season on fecal coliform concentrations in and around vegetative filter strip (VFS) system. Values are reported as log 10 transformation of CFU per 100 mL of sample.

Site a,b,c

Stream

Season d,e,f,g

Manifold

1

2

3

4

5

Control

SEM h

Spring

5.80

2.22

1.27

2.89

2.69

3.46

2.41

0.45

Summer

6.80

2.84

3.03

3.07

3.18

4.76

3.04

0.47

Fall

6.85

4.40

4.32

3.65

3.87

5.68

3.17

0.56

Winter

4.40

2.43

1.22

1.33

0.76

3.96

1.79

0.76

a 1 = Northwest corner of VFS, 2 = Middle north edge of VFS, 3 = North of Northeast corner of VFS, 4 = East of Northeast corner of VFS, 5 = East edge of VFS, Control = Stream site 1 mile south of VFS

b Site effect, P < 0.0001.

c Single df orthogonal contrasts: Manifold vs. Stream, P < 0.0001; Site 5 vs. Other stream sites, P < 0.0001; Site 2 vs. Site 5, P < 0.0001; Manifold vs. Site 5, P = 0.0003; Control stream vs VFS stream, NS.

d Spring = April 1-June 30, Summer = July 1-September 30, Fall = October 1-November 30, Winter = December 1-March 31.

e Season effect, P = 0.0002.

f Site by season interaction, P = 0.005.

g Single df orthogonal contrasts: Spring & Summer vs. Fall & Winter, NS; Summer & Fall vs. Winter & Spring, P < 0.0001.

h SEM = Standard Error of the Mean

Table 2. Treatment least squares means for the effects of sampling site and season on fecal streptococci concentrations in and around vegetative filter strip (VFS) system. Values are reported as log 10 transformation of CFU per 100 mL of sample.

Site a,b,c

Stream

Season d,e,f,g

Manifold

1

2

3

4

5

Control

SEM h

Spring

7.41

2.49

2.09

2.31

2.32

3.73

2.83

0.45

Summer

7.54

3.95

4.07

3.24

2.96

5.79

3.37

0.48

Fall

8.13

4.88

4.62

3.54

3.59

6.52

3.75

0.56

Winter

6.31

3.14

2.31

3.98

4.18

4.54

2.37

0.76

a 1 = Northwest corner of VFS, 2 = Middle north edge of VFS, 3 = North of Northeast corner of VFS, 4 = East of Northeast corner of VFS, 5 = East edge of VFS, Control = Stream site 1 mile south of VFS

b Site effect, P < 0.0001.

c Single df orthogonal contrasts: Manifold vs. Stream, P < 0.0001; Site 5 vs. Other stream sites, P < 0.0001; Site 2 vs. Site 5, P < 0.0001; Manifold vs. Site 5, P < 0.0001; Control stream vs VFS stream, P = 0.006.

d Spring = April 1-June 30, Summer = July 1-September 30, Fall = October 1-November 30, Winter = December 1-March 31.

e Season effect, P = 0.0003.

f Site by season interaction, P = 0.0005.

g Single df orthogonal contrasts: Spring & Summer vs. Fall & Winter, NS; Summer & Fall vs. Winter & Spring, P < 0.0001.

h SEM = Standard Error of the Mean

Table 3. Treatment least squares means for the effect of season on antibiotic resistance of fecal streptococci. Values are expressed as 0 = susceptible, 1 = intermediate resistance, and 2 = resistant.

Level of Significance ( P )

Season a

Contrast b

Antibiotic

Spring

Summer

Fall

Winter

SEM c

Season

1

2

Ciprofloxacin

0.11

0.27

0.04

0.20

0.03

0.007

NS d

NS

Vancomycin

0.03

0.05

0.00

0.02

0.02

NS

NS

NS

Erythromycin

0.21

0.30

0.11

0.15

0.03

0.02

0.04

NS

Streptomycin

1.02

1.17

0.85

0.96

0.06

0.02

0.0005

NS

Neomycin

0.78

0.96

0.64

0.85

0.04

0.01

0.02

NS

Tetracycline

0.37

0.23

0.24

0.48

0.04

NS

NS

0.002

Oxytetracycline

0.38

0.24

0.23

0.53

0.04

0.05

NS

0.0007

a Spring = April 1-June 30, Summer = July 1-September 30, Fall = October 1-November 30, Winter = December 1-March 31.

b Single degree of freedom orthogonal contrasts: 1 = Spring & summer vs. Fall & winter, 2 = Summer & fall vs. Winter & spring.

c SEM = Standard Error of the Mean

d NS = P > 0.05.

Table 4. Treatment least squares means for the effects of sampling site on antibiotic resistance of fecal streptococci. Values are expressed as 0 = susceptible, 1 = intermediate resistance, and 2 = resistant.

Site

Level of Significance ( P )

Stream a

Contrast b

Antibiotic

Manifold

1

2

3

4

5

Control

SEM c

Site

1

2

3

4

5

Ciprofloxacin

0.29

0.10

0.16

0.11

0.14

0.19

0.11

0.04

NS d

NS

NS

NS

NS

NS

Vancomycin

0.12

0.00

0.03

0.00

0.03

0.04

0.00

0.02

NS

NS

NS

NS

NS

NS

Erythromycin

0.27

0.16

0.27

0.17

0.17

0.18

0.14

0.05

NS

NS

NS

NS

NS

NS

Streptomycin

0.94

0.92

1.06

0.98

1.18

1.03

0.89

0.08

0.03

NS

NS

0.003

0.005

NS

Neomycin

0.83

0.69

0.88

0.76

0.91

0.79

0.77

0.06

NS

NS

NS

NS

NS

NS

Tetracycline

0.33

0.37

0.30

0.31

0.35

0.44

0.21

0.06

NS

0.009

0.0008

0.01

NS

NS

Oxytetracycline

0.38

0.39

0.30

0.32

0.38

0.40

0.25

0.06

NS

0.005

0.003

0.01

0.05

NS

a 1 = Northwest corner of VFS, 2 = Middle north edge of VFS, 3 = North of Northeast corner of VFS, 4 = East of Northeast corner of VFS, 5 = East edge of VFS, Control = Stream site 1 mile south of VFS

b 1 = Manifold vs. Stream; 2 = Site 2 vs. Site 5; 3 = Control vs. Stream sites; 4 = Control vs. Sites 1, 2, 3, 4; 5 = Manifold vs. Site 5.

c SEM = Standard Error of the Mean

d NS = P > 0.05.

Table 5. Discriminant analysis of antibiotic resistance patterns of fecal streptococci isolated from animal samples and from water samples.

Number (%) of isolates classified as:

Source

Cow

Deer

Pigeon

Total

Cows

246 (96)

5 (2)

6 (2)

257 (100)

Deer

1 (2)

47 (96)

1 (2)

49 (100)

Pigeons

0 (0)

18 (23)

61 (77)

79 (100)

Total for animal isolates

247 (64)

70 (18)

68 (18)

385 (100)

Manifold

151 (90)

14 (8)

4 (2)

169 (100)

VFS Stream Sites

760 (87)

90 (10)

27 (3)

877 (100)

Control Stream Site

177 (89)

17 (8)

5 (3)

199 (100)

Total for water isolates

1088 (87)

121 (10)

36 (3)

1245 (100)

Table 6. Treatment least squares means for the effect of species on antibiotic resistance of fecal streptococci. Values are expressed as 0 = susceptible, 1 = intermediate resistance, and 2 = resistant.

Level of Significance ( P )

Species

Contrast a

Antibiotic

Cow

Deer

Pigeon

SEM b

Species

1

2

3

4

5

Ciprofloxacin

0.11

0.03

0.02

0.03

0.02

NS c

0.01

NS

NS

NS

Vancomycin

0.02

0.00

0.01

0.01

NS

NS

NS

NS

NS

NS

Erythromycin

0.11

0.06

0.13

0.04

NS

NS

NS

NS

NS

NS

Streptomycin

1.41

0.84

1.32

0.09

<0.0001

<0.0001

NS

0.001

<0.0001

NS

Neomycin

0.91

0.58

0.71

0.06

0.0008

0.0007

0.01

NS

0.02

NS

Tetracycline

0.63

0.00

0.79

0.09

<0.0001

<0.0001

NS

<0.0001

<0.0001

<0.0001

Oxytetracycline

0.66

0.00

0.76

0.09

<0.0001

<0.0001

NS

<0.0001

<0.0001

0.0002

a Single degree of freedom orthogonal contrasts: 1 = Cow vs Deer, 2 = Cow vs Pigeon, 3 = Pigeon vs Deer, 4 = Cow & Pigeon vs Deer, 5 = Cow & Deer vs Pigeon.

b SEM = Standard Error of the Mean

c NS = P > 0.05.