th Biannual Meeting of the Hellenic (Greek) Society for Basic & Clinical Pharmacology Professor Arthur Christopoulos, B.Pharm., Ph.D., Drug Discovery, Monash University, Australia Today's science, tomorrow's medicines Athens 23-24 RegistrationYoung Investigators ForumChairs: Dr. E. Papadimitriou
976-981.qxdBacteremia associated with naturally occurring
acute coliform mastitis in dairy cows
John R. Wenz, DVM, MS; George M. Barrington, DVM, PhD, DACVIM; Franklyn B. Garry, DVM, MS, DACVIM; Kevin D. McSweeney, BS; R. Page Dinsmore, DVM, DABVP; Gregory Goodell, DVM; Robert J. Callan, DVM, PhD, DACVIM become the predominant form of mastitis in herds inwhich contagious mastitis has been effectively con- Objective—To determine the incidence of bacteremia
trolled.1-3 Despite decades of research focused on ACM, in dairy cows with naturally occurring acute coliform wholly effective control measures have yet to be estab- mastitis (ACM) with a wide range of disease severity.
lished, and proper treatment of the disease remains controversial. Our understanding of ACM pathophysi- Animals—144 dairy cows with ACM from 6 herds.
ology is based largely on studies of the experimental Procedure—Cows were examined at time of identifi-
disease model. These studies typically involved inocu- cation of ACM (time 0) and classified as having mild, lation of a healthy mammary gland with coliform moderate, or severe mastitis on the basis of rectal organisms (Aerobacter aerogenes, E coli, and Klebsiella temperature, hydration status, rumen contraction rate, spp) or with purified endotoxin. Results of these and attitude. Cows were reexamined at 24 or 48 hours.
experiments suggest coliform bacteria are noninvasive, Bacteriologic culturing of milk and blood (30 ml), CBC, and bacteremia (the presence of bacteria in the circu- and serum biochemical analysis were performed at lating blood) is not considered a significant sequela to each time point. Appropriate samples were obtained at ACM.4-6 Furthermore, disease manifestation has been a single point from herdmates without mastitis (con- primarily attributed to the effects of endotoxin.
trols) that were closely matched for lactation number Consequently, it has been widely held that antibiotic and days since parturition. Blood culture results werecompared among severity groups and controls by use therapy is not warranted in cases of ACM. We feel that of χ2 tests, as was outcome of an ACM episode for studies of naturally occurring ACM will help resolve cows grouped by blood bacterial isolates.
discrepancies between knowledge based on the exper- Results—Bacteria were isolated from 52 blood samples
imental disease model and practitioner experience in from 46 of 144 (32%) cows with ACM, which was signif- icantly more than control cows (11/156; 7.1%). Group-1 In 1 study7 of 20 cases of naturally occurring ACM, isolates (Escherichia coli, Pasteurella multocida, only Bacillus spp were isolated from a few cases, a find- Mannheimia haemolytica, Klebsiella pneumoniae, ing that was attributed to skin contamination. The Enterobacter agglomerans, and Salmonella enterica severity of disease was not clearly defined, only one 5- serotype Typhimurium) were identified in 20 of 144 (14%) ml blood sample was evaluated, and controls were not cows with ACM and 0 of 156 control cows. Group-1 iso- included in the study. In our previous work,8 the exis- lates were identified in 4.3, 9.1, and 42% of cows classi- tence of bacteremia associated with naturally occurring fied as having mild, moderate, and severe ACM, respec-tively. Escherichia coli and K pneumoniae milk and blood ACM was established, in which bacteremia associated isolates obtained from the same cow were of the same with E coli was identified in 32% (11/34) of cows with genotype. Bacillus spp were identified in 21 of 144 (15%) severe protracted coliform mastitis. That study, howev- cows with ACM, which was significantly more than con- er, lacked controls and was limited to a select popula- trol cows (3/156; 1.9%). Thirty-five percent of cows with tion of cows with severe protracted disease.
a group-1 isolate died during the mastitis episode.
Furthermore, it has been suggested that bacteremia Conclusions and Clinical Relevance—Results sug-
observed in these cows may have been the result of pro- gest that bacteremia develops in a substantial pro- tracted disease rather than specific to the ACM portion of cows with ACM. Classification of severity episode.6 Recently, we have reported important patho- of disease is important for establishment of effective physiologic differences among cows with ACM grouped treatment protocols; parenteral antimicrobial treat- by severity of systemic disease signs.9 ment may be indicated in cows with ACM. (J Am Vet The purposes of the study reported here were to Med Assoc 2001;219:976–981) determine the incidence of bacteremia in dairy cows Acute coliform mastitis (ACM), typified by with naturally occurring ACM and a wide range of dis-
Escherichia coli intramammary infection, has ease severity, determine whether the affected mamma-ry gland was the source of bacteremia, and identify fac- From the Department of Clinical Sciences, College of Veterinary tors that may be useful in predicting the occurrence of Medicine, Colorado State University, Ft. Collins, CO 80523. Dr.
bacteremia in dairy cows with ACM.
Barrington's present address is Washington State University,Department of Veterinary Clinical Sciences, College of Veterinary Materials and Methods
Medicine, Pullman, WA 99164.
Supported by the Colorado Agricultural Experiment Station, Herds—Cows at 6 dairies surrounding Fort Collins,
Pharmacia Animal Health, and The American Association of Colo, that developed clinical mastitis between July 1997 and January 1999 were eligible for inclusion in the study. During The authors thank Jane Carmen for technical assistance.
the study, there were approximately 4,000 Holstein cows in Scientific Reports: Original Study JAVMA, Vol 219, No. 7, October 1, 2001 lactation at the 6 dairies. Cows were fed a total mixed ration
agar plates. Plates were incubated at 37 C in an atmosphere (TMR) in groups based on production level and were milked
containing 10% CO2 and examined after 24 and 48 hours.
in a parlor 3 times daily. All cows were housed in drylot pens Bacterial colonies were identified in accordance with or freestalls. Mean somatic cell count of bulk tank milk was National Mastitis Council guidelines.10 Coliform intramam- < 300,000 cells/ml throughout the study period for all dairies.
mary infection was diagnosed if samples contained ≥ 10 All cows were vaccinated with a bacterin containing the J-5 colony-forming units (cfu) of at least 1 coliform organ-
strain of E colia at the start of the nonlactating period and 4 ism/ml. Plates with ≥ 3 organisms were considered contami- weeks before and within 2 weeks after parturition. Heifers nated. Bacterial numbers obtained following culture were received the J-5 bacterin 2 weeks before and within 2 weeks reported as < 10,000, 10,000 to 100,000, and > 100,000 after parturition. All dairies practiced effective contagious mastitis control programs and participated in a program for Blood was collected from the jugular vein of cows with monthly bacterial culture of bulk tank milk samples.
ACM at time 0 and 24 hours later or at time 0 and 48 hourslater and submitted for bacterial culture. Blood from control Inclusion criteria and data collection—Farm personnel
cows was collected at a single time point. Hair was shaved, identified cows as possibly having coliform mastitis if 1 or using a disposable razor, and the skin was disinfected with at more of the following signs were present: reduced milk pro- least 3 alternating applications of povidone iodine scrub and duction, abnormal milk, and 1 or more abnormal mammary 70% ethanol at a site over the jugular vein. Thirty milliliters quarters. Initial examination and sample collection (time 0) of blood was aseptically drawn from the vein into a 35-ml was performed by 1 of the authors (JRW, KM), usually with- syringe through an 18-ga needle. Fifteen milliliters of blood in an hour after notification from the farm. Cows identified was injected aseptically through a new 18-ga needle into each with mastitis during the evening or night milking were exam- of two 50-ml blood culture vials of brain-heart infusion broth ined the following morning. Farm personnel treated cows containing 0.6% sodium polyanetholsulfonate.b Samples were according to established farm protocols after initial examina- aerated through a filtered needle and incubated at 37 C in an tion and sample collection by study personnel. Treatment atmosphere containing 10% CO2. Samples were subcultured was not controlled in order to maximize producer participa- onto blood agar on days 0, 1, and 7 and incubated at 37 C in tion in the study and determine the incidence of bacteremia an atmosphere containing 10% CO2. Plates were examined for associated with ACM given current varied farm treatment growth and recorded at 24 and 48 hours. Escherichia coli, K protocols. Treatment among farms was consistent in the use pneumoniae, Pasteurella multocida, Mannheimia haemolytica, of anti-inflammatory drugs (flunixin meglumine, phenlybu- tazone), 7.2% sodium chloride solution (administered IV), Typhimurium isolates were considered to be important path- and electrolyte solutions (administered PO). Antimicrobial ogenic bacteria and were classified as group-1 isolates.
therapy varied among and within farms and included com- Environmental streptococci, coagulase-negative staphylococ- mercially available antibiotic preparations (intramammary), ci, and Acinetobacter spp were considered to be coincidental gentamicin (intramammary), dexamethasone (intramamma- findings and classified as group-2 isolates. Bacillus spp were ry), ceftiofur (IV and IM), and oxytetracycline (IV).
Secretions were removed from affected mammary glands 3times a day at the time of milking and an additional 2 times Coliform isolate genotyping—Genotyping of E coli and
a day during treatments. Cows for which bacterial culture of K pneumoniae was performed on isolates obtained from milk a milk sample yielded E coli or Klebsiella spp were included and blood of the same cow as described previously.11 in the study. Healthy herdmates, closely matched to infected Enterogenic repetitive intergenic consensus sequence cows by lactation number and days since parturition, were primers were used to perform polymerase chain reaction
enrolled as controls. (PCR) on the genomic DNA of E coli and K pneumoniae iso-
Cows were evaluated at the time of initial examination lates. The presence and size (base pairs) of PCR products (time 0) and again 24 hours later or at time 0 and again 48 were compared between milk and blood coliform isolates hours later. Control cows were examined at 1 of the time obtained from the same cow.
points. Age, lactation number, days since parturition, rectaltemperature, heart rate, respiratory rate, and rumen contrac- Hematologic testing—Blood samples were collected
tion rate were recorded. Hydration status was estimated on from the tail vein at time 0 and 24 hours later or time 0 and 48 the basis of degree of enophthalmos and scored as 0 (none), hours later. Testing of blood from control cows was not per- 1 (mild enophthalmos), 2 (moderate enophthalmos), or 3 formed, because laboratory reference ranges were already (severe enophthalmos). Attitude was classified on the basis established. Samples anticoagulated with EDTA were used for of signs of depression as none, mild, or severe. CBC and determination of plasma protein and fibrinogen con-centrations. Samples collected in plain tubes without anticoag- Severity classification—Cows with ACM were classi-
ulant were used to harvest serum, and serum glucose, creati- fied as having mild, moderate, or severe disease on the basis nine, total protein, albumin, globulin, total bilirubin, phos- of rectal temperature, hydration status, rumen contraction phorus, calcium, magnesium, sodium, potassium, chloride, rate, and attitude at time 0 (Appendix).
and bicarbonate concentrations; aspartate transaminase, γ-glu-tamyltransferase, sorbitol dehydrogenase, and creatine kinase Bacteriologic culture—Secretions from affected mam-
activities; and anion gap were determined. Hematologic values mary glands were collected at time 0 and 24 hours later or were compared between cows with a group-1 bacteremia and time 0 and 48 hours later and submitted for bacterial culture.
those in which blood culture results were negative or a A composite sample of milk from all quarters of control cows non–group-1 bacteremia within each severity group.
was collected at a single time point. Teat ends were disinfect-ed with 70% ethanol prior to sample collection, and samples Outcome—Survival and retention in the herd following
were collected in sterile vials. Samples were stored on ice for the episode of mastitis were assessed to determine outcome.
transport to the laboratory and processed the day of collec- Cows that died during the episode of mastitis were classified tion, except that those submitted to the laboratory after 5:00 as nonsurvivors; all others were considered survivors. Cows PM were frozen at –4 C and processed the following day.
leaving the herd as a direct result of mastitis (ie, because of Samples were mixed gently by inverting the tube, and 100 µl mastitis or low production) within 30 days after the episode of each sample was plated on blood agar and MacConkey of mastitis were classified as culled. Cows still present in the JAVMA, Vol 219, No. 7, October 1, 2001 Scientific Reports: Original Study herd 30 days after the episode of mastitis were classified as single blood sample from 8 cows and in combination with Bacillus spp in 4 cows with ACM. Bacillus spp Data analyses—Continuous variables (days since partu-
alone were identified in a single blood sample of 14 cows rition, age, lactation, and results of clinicopathologic tests) with ACM. Bacteria were isolated from 30, 8, and 14 were tested for normality by use of the Kolmogorov D statis- blood samples of cows with ACM taken at 0, 24, and 48 tic. Normally distributed continuous variables were com- hours, respectively. Bacteria were isolated from the pared among groups (mild, moderate, or severe disease) by blood of cows with ACM at both 0 and 24 hours from 3 use of ANOVA. Continuous variables that were not normally cows and at both 0 and 48 hours from 3 cows. Group-1 distributed were compared among groups by use of isolates were identified at time 0 in 11 cows, at 24 hours Wilcoxon rank sum or Kruskal-Wallis tests. Categorical vari- in 3 cows, at 48 hours in 4 cows, at 0 and 24 hours in 1 ables (milk bacterial count, proportion of bacteremic cows) cow, and at 0 and 48 hours in 1 cow. Cows were classi- were compared among groups by use of χ2 tests, except thatFisher exact tests were used when the expected count in fied as bacteremic if bacteria were isolated from at least > 25% of the categories was < 5. Influence of farm on blood 1 blood sample at any time point. Group-1 bacteremia culture results and outcome data was evaluated by use of a was identified in a significantly (P < 0.001) greater per- maximum-likelihood ANOVA procedure.c Statistical calcula- centage of cows with ACM (14%; 20/144), compared tions were performed by use of commercially available soft- with control cows (0%; 0/156). Fifteen percent (21/144) ware.c Relative risk was calculated as an odds ratio; the odds of cows with ACM had a Bacillus spp bacteremia, which ratio was not considered significantly different from 1 if its was significantly (P < 0.001) more than control cows 95% confidence interval included 1. For all other tests, val- (1.9%; 3/156). The percentage of cows with group-2 iso- ues of P < 0.05 were considered significant.
lates was not significantly (P = 0.2) different betweencows with ACM and controls.
Samples were collected from 178 cows with sus- When evaluated on the basis of severity of ACM, a pected ACM. Cows were excluded if samples yielded significantly greater percentage of cows in the severe no growth on 0-hour milk culture (n = 13), a noncol- group had a group-1 bacteremia (42%), compared with iform organism was isolated (n = 15), or the 0-hour moderate (9.1%) and mild (4.3%) groups (Table 1). In
milk sample was contaminated (n = 2). Four cows were contrast, significant differences were not detected in excluded because data necessary for severity classifica- the percentage of cows with group-2 or Bacillus spp tion were missing.
isolates among severity groups. The odds of a group-1bacteremia in cows in the severe group were 15 (95% Bacteriologic testing—Coliform organisms were
confidence interval [CI], 4.1 to 61.9) and 7.2 (95% CI,
isolated from the time-0 milk culture of 144 cows that 2.1 to 25.2) times greater than in cows in the mild and were subsequently enrolled in the study. Organisms moderate groups, respectively. included E coli (n = 122), K pneumoniae (n = 19), and Group-1 bacteremia was identified in a significant- 3 cases in which both E coli and K pneumoniae were ly greater percentage of cows with > 100,000 cfu/ml in milk from the infected gland at time 0 (23%; 15/64) Bacteria were isolated from 52 blood samples in 46 than in cows with < 100,000 cfu/ml (5.3%; 4/76).
of 144 (32%) cows with ACM, which was significantly Significant differences in cfu/ml bacteria in milk from (P < 0.001) more than those isolated from 11 of 156 the infected gland were not detected in cows with (7.1%) control cows. Group-1 isolates (E coli, K pneu- group-2 or Bacillus spp isolates alone.
moniae, Pasteurella multocida, Mannheimia haemolytica,E agglomerans, and Salmonella Typhimurium), group-2 Coliform isolate genotyping—Genotyping was
isolates (environmental streptococci, coagulase nega- performed on coliform organisms of the same genus tive staphylococci, and Acinetobacter spp), and Bacillus isolated from the milk and blood of the same cow.
spp isolates were evaluated independently. Group-1 Escherichia coli was isolated from the milk and blood of isolates alone were identified in a single blood sample 10 cows, and K pneumoniae was isolated from the milk from 14 cows, in combination with a group-2 isolate in and blood of 1 cow. The DNA fingerprint of E coli milk 2 cows, and in combination with Bacillus spp in 4 cows and blood isolates, obtained from the same cow, had with ACM. Group-2 isolates alone were identified in a the same banding pattern, indicating they were the Table 1—Number and percentage of positive blood culture results by bacterial isolate from cows withacute coliform mastitis classified by severity of disease and control cows Group-1 Group-2 Bacillus
Values within a column with different superscripts are significantly (P ⬍ 0.05) different. Blood bacterial isolates were eval- uated independently; therefore, cows with more than 1 isolate type were counted more than once.
Group-1 isolates = Escherichia coli, Pasteurella multocida, Mannheimia hemolytica, Klebsiella pneumoniae, Enterobacter agglomerans, Salmonella Typhimurium. Group-2 isolates = Environmental streptococci, coagulase-negative staphylococci,and Acinetobacter spp. Scientific Reports: Original Study JAVMA, Vol 219, No. 7, October 1, 2001 Table 2—Survival and retention rates in a group of 144 cows with acute coliform mastitis Negative Group-1 Group-2 Bacillus
No. % No. % No. % No. %
*Values were significantly (P ⬍ 0.001) different. Cows were grouped according to results of bacteriologic culture of blood.
Comparisons were only made between groups of isolates and cows with negative culture results; comparisons were notmade between bacteremic groups. Died = No. of cows that died during the mastitic episode. Retained = No. of cows still present in the herd 30 days after the mastitic episode. Culled = No. of cows removed from the herd as a direct result of mastitis.
See Table 1 for remainder of key.
same genotype. The K pneumoniae milk and blood iso- tured (Table 1). Antimicrobial use was variable among lates from the same cow were also of the same genotype.
and within farms and may have had an impact onresults of this study. Therefore, the proportion of bac- Differentiation of group-1 bacteremic cows and
teremic cows may be higher in cases of ACM that are cows with negative blood culture results or
not treated with antibiotics. non–group-1 bacteremia—Signalment and hematologic
Intermittent bacteremia is commonly the result of data were compared between cows with group-1 bac- bacterial infiltration of the blood through the lymphatics teremia and those with negative blood culture results or from a site of localized infection.12 It has been suggested non–group-1 bacteremia within each severity group.
that the E coli bacteremia observed by Cebra et al8 may Significant differences were not detected in mean number have been nonspecific and attributable to prolonged of days since parturition, age, and lactation number of severe disease. Indeed, E coli isolated from the blood cows with group-1 bacteremia versus those without by samples in that study could have come from the large severity group.9 Complete hematologic data were available pool in the gastrointestinal tract. However, antibiotic sen- for 141 cows. Median values of CBC and serum biochem- sitivity patterns of E coli isolated from milk and blood of ical analysis by severity group have been reported. 9 There the same cow were similar, which suggests a common were no significant differences in hematologic values of source. Furthermore, in the present study, genotyping by cows with group-1 bacteremia versus those without by use of PCR indicated that E coli and Klebsiella spp isolat- severity group. Significant and clinically important differ- ed from blood were the same as those isolated from the ences did exist in the severity classification of cows with mammary gland of the same cow, indicating the infected group-1 bacteremia versus those without (Appendix).
mammary gland was the likely source of bacteremia.
Sixty-five percent (13/20) of cows with group-1 bac- Resident macrophages in the liver and spleen are teremia were classified as severe, compared with only 15% primarily responsible for bacterial clearance of the (18/122) of those without group-1 bacteremia.
blood; however, bacterial opsonization and subsequent Outcome—Outcome was evaluated based on sur-
phagocytosis by circulating leukocytes play a role as vival through the mastitic episode and retention on well.13 Leukopenia, specifically neutropenia, has been farm for 30 days following the episode (Table 2). No
well documented in cows with ACM. Furthermore, significant differences in the percentage of cows that results from our previous work revealed that cows with died or were culled were observed between cows with more severe systemic disease signs have more pro- a group-2 isolate or Bacillus spp isolate versus cows found neutropenia,9 which may increase the probabili- with negative blood culture results. A significantly (P < ty of obtaining positive blood culture results in cows 0.001) greater percentage (35%) of cows with a group- with more severe ACM. Indeed, cohort studies of 1 isolate died, compared with cows with negative immunocompromised humans have demonstrated a blood culture results or a non-group 1 isolate (0%).
positive relationship between neutropenia and bac- Significant differences in culling were not observed teremia attributable to gram-negative bacilli such as E among cows grouped by blood culture results. coli and Klebsiella spp.14 However, although it wasreported that a significant difference was not detected in median neutrophil count in cows with moderate Results of this study indicated bacteremia develops versus severe systemic disease signs,9 results of the pre- in a substantial proportion of cows with naturally sent study indicated there were significantly more pos- occurring ACM. Furthermore, bacteremia caused by itive blood culture results obtained from cows with group-1 pathogens (E coli, P multocida, M haemolytica, severe ACM (48%), compared with cows with moder- K pneumoniae, E agglomerans, and Salmonella ate ACM (23%). Furthermore, despite the profound Typhimurium) develops in a significant proportion of neutropenia observed in cows with moderate ACM, cows with ACM and is associated with severity of sys- compared with those with mild ACM,9 there was no temic disease signs. Our results also revealed that a sig- difference in positive blood culture results between the nificantly greater percentage of cows in the mild and 2 groups (Table 1). The same was true when only moderate groups had group-1 bacteremia, compared group-1 blood isolates were considered. Our results with controls from which no group-1 isolates were cul- suggest that a factor other than neutropenia may be JAVMA, Vol 219, No. 7, October 1, 2001 Scientific Reports: Original Study involved in the pathophysiology of bacteremia associ- severity groups, all cows with Bacillus spp bacteremia ated with ACM.
survived, and there was no difference in cull rate, com- Many factors may have been responsible for failure pared to cows with negative blood culture results.
to identify bacteremia associated with ACM in previous Group-2 isolates (ie, environmental streptococci, studies.5-7 Most used experimental disease models in coagulase-negative staphylococci, and Acinetobacter which the mammary glands of healthy cows were spp) were considered inconsequential, as there was no inoculated with a specified uniform number of col- difference between controls and cows with ACM.
iform bacteria. Disease manifestation is thought to be Furthermore, there was no difference in outcome the result of interaction of host, pathogen, and envi- between cows with a group-2 isolate alone and those ronmental factors. Epidemiologic studies4,15,16 of ACM with negative blood culture results. Group-2 isolates have indicated that E coli is an accidental pathogen may have been contaminants or present in the blood of with no specific virulence factors involved in establish- all cows periodically. ing infection of the mammary gland. Furthermore, Blood cultures were performed at 2 time points to despite the consistency of inoculum size and control of increase the chance of identifying cows with bac- environmental factors afforded by the experimental teremia. The second time point was 24 hours after ini- model, pathophysiologic response varies widely among tial examination (time 0) at the start of the study; how- individual cows.16 Taken together, these results suggest ever, it was later changed to 48 hours. The 48-hour cow factors (ie, immune status, presence of concurrent time point allowed for better evaluation of changes in disease, teat-end condition) may be most important in hematologic data; furthermore, results of our previous determining the pathogenesis of ACM. Therefore, con- study identified bacteremia in cows that were ill for a siderable differences in disease manifestation may exist median of 48 hours.8 Interestingly, 65% (13/20) of in the experimental disease of healthy cows that cows with group-1 positive blood culture results were arguably have intact host defenses versus naturally identified at time 0, and 35% (7/20) were in the mild occurring disease in cows that are likely to have com- or moderate group. This is further evidence that the promised host defenses.
observed bacteremia was not merely the result of Recently we reported that severity classification severe protracted disease. based on systemic disease signs effectively distin- The design of this study did not allow for discrim- guished significant pathophysiologic differences ination of the cause-and-effect relationship between among cows with ACM.9 Previous studies that failed to severity of disease and bacteremia. However, evaluation identify bacteremia did not evaluate or clearly define of the data collected suggests that bacteremia caused by severity of ACM cases. Results of the present study group-1 pathogens had a negative impact on cows with indicated that group-1 bacteremia develops in a signif- ACM. Death during the mastitis episode only occurred icantly smaller proportion of cows with mild and mod- in cows with group-1 bacteremia. Six of 7 cows that erate disease signs, compared to cows with severe dis- died were in the severe group, suggesting severity of an ease signs (Table 1). Consequently, a predominance of ACM episode is an important factor. Severe disease more mildly affected cows combined with a small sam- alone, however, does not appear to determine survival.
ple size may have resulted in failure to identify bac- Seven of 13 severely affected cows with group-1 bac- teremia in previous studies.
teremia survived, compared with 18 of 18 severely Culture technique may account for failure to iden- affected cows without group-1 bacteremia. Together, tify bacteremia associated with ACM. Previous studies these data suggest group-1 bacteremia was an added typically performed blood culture on 5 ml of blood col- insult to severe disease that resulted in a poorer out- lected at a single time point. Often < 10 bacteria are come. Severe disease, however, is not a prerequisite for present per milliliter of blood in intermittently bac- development of group-1 bacteremia, as 35% of group-1 teremic human patients.17 The chances of obtaining a bacteremic cows were in the mild and moderate groups.
positive culture are directly related to the volume of However, of cows with mild and moderate disease blood collected.18 Increasing the blood volume collect- severity, only 1 of 7 (a cow in the moderate group) with ed from 2 ml to 20 ml resulted in an increase in posi- group-1 bacteremia died.
tive cultures from 30% to 50% in human patients,18-20 The culling rate of surviving cows was the same in whereas volumes > 30 ml did not result in a significant cows with ACM regardless of blood culture results increase in positive yield.20 When sequential blood cul- (Table 2). Likewise, of the cows with severe disease tures were performed in humans that did not have that survived, there was no difference in culling endocarditis, 80% were positive on first culture, 90% between cows with group-1 bacteremia (57%; 4/7) and after second culture, and 99% after third culture.18 In those without (61%; 11/18). Therefore, it appears that the present study, 30-ml blood samples were collected group-1 bacteremia has a negative impact on survival at 2 time points in an attempt to optimize identifica- but not culling during the first 30 days following a tion of bacteremia. mastitis episode.
Bacillus spp have been isolated from blood in pre- Pasteurella multocida or Mannheimia haemolytica vious studies but were presumed to be skin contami- was isolated from the blood of 7 cows, E agglomerans nants.6-8 Inclusion of controls in the present study indi- from 1 cow, and Salmonella Typhimurium from 1 cow cates Bacillus spp bacteremia is 8 times more likely in with ACM. These bacteria were not, however, isolated cows with ACM, compared with controls. The signifi- on milk culture of the same cow. These findings sug- cance and source of Bacillus spp bacteremia in cows gest some cases of bacteremia associated with coliform with ACM is unclear. The incidence was similar among mastitis may be the result of bacterial translocation Scientific Reports: Original Study JAVMA, Vol 219, No. 7, October 1, 2001 from other organ systems. Cows with Pasteurella or acute coliform mastitis in Holstein cattle. J Vet Intern Med 1996; Mannheimia spp bacteremia had more severe disease signs, and their prognosis was grave. These bacteria are 9. Wenz JR, Barrington GM, Garry FB, et al. Use of systemic disease signs to assess severity in dairy cows with acute coliform normal flora of the upper portion of the respiratory tract mastitis. J Am Vet Med Assoc 2001;218:567–572.
of cattle and are involved in the bovine respiratory dis- 10. National Mastitis Council. Laboratory handbook on bovine ease complex. Studies in sheep and in vitro studies of mastitis. Madison, Wis: National Mastitis Council Inc, 1999;31–39.
bovine pulmonary endothelial monolayers indicate that 11. Lipman LJA, de Nijs A, Lam TJGM, et al. Identification of endotoxin can cause direct dose-dependent damage Escherichia coli strains from cows with clinical mastitis by serotyping resulting in an increase in permeability and hydraulic and DNA polymorphism patterns with REP and ERIC primers. VetMicrobiol 1995;43:13–19.
conductance across pulmonary endothelium.21 Such 12. Dow S, Jones R. Bacteremia: pathogenesis and diagnosis.
damage may develop in cows with more severe systemic Compend Contin Educ Pract Vet 1989;11:432–444.
disease signs and result in bacteremia. It has been pro- 13. Nostrandt AC. Bacteremia and septicemia in small animal posed that cows with ACM are not endotoxemic.22 patients. Probl Vet Med 1990;2:348–360.
However, endotoxin was identified in the milk vein 14. Mathews WC, Caperna J, Toerner JG, et al. Neutropenia is blood of 5 of 9 cows with naturally occurring ACM. 23 a risk factor for gram-negative bacillus bacteremia in human immun-odeficiency virus-infected patients: results of a nested case-control Together with the results of the present study, these data study. Am J Epidemiol 1998;148:1175–1184.
suggest cows with more severe systemic disease signs 15. Frost AJ, Hill AW, Brooker BE. The early pathogenesis of may be endotoxemic. Further study is needed to deter- bovine mastitis due to Escherichia coli. Proc R Soc Lond B Biol Sci mine whether cows with ACM are endotoxemic and identify the relationship between possible endotoxemia, 16. Jones TO. Escherichia coli mastitis in dairy cattle—a review bacteremia, and severity of an ACM episode.
of the literature. Vet Bull 1990;60:205–231.
17. Tilton R. The laboratory approach to the detection of bac- Signalment and hematologic data were not useful in teremia. Annu Rev Microbiol 1982;36:467–493.
identifying a cow with group-1 bacteremia. Classification 18. Washington J. Blood cultures: Principles and techniques.
of disease severity based on systemic disease signs was Mayo Clin Proc 1975;50:91–98.
the best indicator of a cow likely to be bacteremic.
19. Hall MM, Ilstrup DM, Washington JA. Effects of volume of The results of this study clearly demonstrate that blood cultured on detection of bacteremia. J Clin Microbiol bacteremia develops in a substantial proportion of 20. Tenney J, Reller L, Mirrett S, et al. Controlled evaluation of cows with ACM and has an impact on the outcome of the volume of blood cultured in detection of bacteremia and an ACM episode. We also demonstrated the impor- fungemia. J Clin Microbiol 1982;15:558–562.
tance of accurate classification of disease severity 21. Meyrick B. Endotoxin-mediated pulmonary endothelial cell based on systemic disease signs. Classification of dis- injury. Fed Proc 1986;45:19–24.
ease severity should play an important role in estab- 22. Lohuis J, Van Leeuwen W, Verheijden J, et al. Effect of dex- lishing rational effective treatment protocols for cows amethasone on experimental Escherichia coli mastitis in the cow. JDairy Sci 1988;71:2782–2789.
with ACM. On the basis of results of this study, there 23. Katholm J, Andersen P. Acute coliform mastitis in dairy should be a high index of suspicion of bacteremia in cows: endotoxin and biochemical changes in plasma and colony- cows with severe systemic disease signs; parenteral forming units in milk. Vet Rec 1992;131:513–514.
antimicrobial therapy may be indicated in such casesof ACM.
Scheme based on systemic disease signs for classifying sever-
aJ-5 Bacterin, Pharmacia and Upjohn Animal Health, Kalamazoo, ity of acute coliform mastitis in dairy cows* bBBL, Becton Dickinson Co, Cockeysville, Md.
cSAS/STAT, release 6.12, SAS Institute Inc, Cary, NC.
Rectal temperature 37.8 (100)–39.27 (102.7) 39.33 (102.8)–39.8 (103.7) 1. Erskine R, Eberhart R, Hutchinson L, et al. Incidence and ⬎ 39.8 (103.7) or ⬍ 37.8 (100) types of clinical mastitis in dairy herds with high and low somaticcell counts. J Am Vet Med Assoc 1988;192:761–765.
Hydration status 2. Gonzalez RN, Jasper DE, Kronlund NC, et al. Clinical mas- (degree of enophthalmos) titis in two California dairy herds participating in contagious masti- tis control programs. J Dairy Sci 1990;73:648–660.
3. Fox L. Introduction to coliform mastitis, in Proceedings.
Coliform Mastitis Symp 1993;7–9.
Rumen contraction rate 4. Frost AJ, Hill AW, Brooker BE. Pathogenesis of experimen- (contractions/min) tal bovine mastitis following a small inoculum of Escherichia coli. Res Vet Sci 1982;33:105–112.
5. Pyorala S, Kaartinen L, Kack H. Efficacy of two therapy reg- imens for treatment of experimentally induced Escherichia coli mas- (signs of depression) titis in cows. J Dairy Sci 1994;77:453–461.
6. Shpigel NY. Should we use antimicrobials for treatment of coliform mastitis in dairy cows? Cattle Pract 1998;6:113–120 7. Powers M, White ME, Dinsmore RP, et al. Aerobic blood *Cows with total score of 0 to 2 were classified as having mild disease, culturing in cows with coliform mastitis. J Am Vet Med Assoc 1986; cows with total score of 3 to 5 were classified as having moderate disease, and cows with total score of 6 to 9 were classified as having severe disease.
8. Cebra CK, Garry FB, Dinsmore RP. Naturally occurring JAVMA, Vol 219, No. 7, October 1, 2001 Scientific Reports: Original Study
Contents lists available at Artiﬁcial Intelligence in Medicine A semantic graph-based approach to biomedical summarisation Laura Plaza , Alberto Díaz, Pablo Gervás Departamento de Ingeniería del Software e Inteligencia Artiﬁcial, Universidad Complutense de Madrid, C/Profesor José García Santesmases, s/n, 28040 Madrid, Spain Objective: Access to the vast body of research literature that is available in biomedicine and related