Effect of Prepartum Propylene Glycol Administrationon Perlparturient Fatty Liver In Dairy Cows

VAUGHN A. STUDER, RIC R. GRUMMER,1 and SANDRA J. BERTICSDepartment of Dairy Science

University of WisconsinMadison 53706

CHRISTOPHER K. REYNOLDSRuminant Nutrition Laboratory

US Department of Agriculture, ARSBeltsville, MD 20705

ABSTRACT

Plasma glucose concentration duringlate gestation was thought to be impor-tant for the development of fatty livernear parturition. Thirteen multiparouscows were given a l-L oral drench ofpropylene glycol once daily beginning10 3.6 d prepartum until parturition.Eleven control cows received a l-L wa-ter drench. Plasma glucose increased fol-lowing propylene glycol administration.Plasma NEFA concentration was 403and 234 p.M, and plasma insulin concen-trations were .354 and .679 ng/ml, forcontrol cows and cows treated withpropylene glycol measured from 1 to 7 dprepartum. Plasma NEFA tended to belower in cows treated with propyleneglycol from 1 to 21 d postpartum.Prepartum propylene glycol administra-tion reduced hepatic triglyceride ac-cumulation by 32 and 42% at 1 and 21 dpostpartum, respectively. Prepartumplasma BHBA was reduced duringpropylene glycol administration. Prepar-tum plasma glucose, NEFA, BHBA, andinsulin were strongly correlated withliver triglyceride at 1 d postpartum (r =-.49, .45, .36, and -.49, respectively).Pre- and postpartum OMI were not af-fected by treatment. Milk production andcomposition measured through 21 dpostpartum were not different betweengroups.(Key words: prepartum, propyleneglycol, fatty liver, dairy cows)

Received March 25, 1993.Accepted May 17. 1993.lReprint requests: 1675 Observatory Drive.

1993 J Dairy Sci 76:2931-2939

Abbreviation key: PG = propylene glycol,RIA =radioimmunoassay, TG =triglyceride.

INTRODUCTION

Fatty liver is a metabolic disorder as-sociated with a decrease in animal production(10), reproductive performance (10), and im-mune competence (19). Decreased hepaticgluconeogenesis has also been associated withconditions conducive to fatty liver develop-ment (12). Recent evidence (31) suggests thatfatty liver may precede clinical ketosis.

During periods of excessive fatty acidmobilization from adipose tissue, ruminantsare prone to the development of fatty liver.Hepatic uptake of NEFA is positively relatedto plasma concentration (1). Once taken up byhepatocytes, fatty acids can either be com-pletely oxidized to carbon dioxide, partiallyoxidized to acetyl-coenzyme A and released asketone bodies or acetate, or esterified.Triglyceride (fG), the main esterification prod-uct, can be stored or exported within very lowdensity lipoprotein. Accumulation of TO inisolated bovine hepatocytes increased asNEFA concentration increased in the media(5). Capacity for ruminants to export very lowdensity lipoproteins is low (16, 23); thus, ex-cessive TO accumulation is likely during rapidNEFA uptake by the liver.

Fatty liver is commonly considered to de-velop postpartum. However, hepatic TO ac-cumulation is significant between 17 d prepar-tum and 1 d postpartum (2, 29). Estrogen,specifically estrone of placental origin, iselevated in late gestation and has been impli-cated in the development of fatty liver (13).Feed intake is depressed immediately prepar-tum (2), and negative energy balance and

2931

2932 STUDER ET AL.

NEFA release from adipose tissue can occureven prior to copious milk secretion. Recently,our laboratory (2) demonstrated that hepaticTG content at 1 d postpartum was significantlyreduced in cows that were force-fed prepartumvia rumen cannulas. Plasma glucose at 2 dprepartum also was elevated in force-fed cowsand was negatively correlated with liver TG atd I postpartum.

Propylene glycol (pG) is a glucogenic com-pound used to treat ketosis postpartum. Themajority of PG escapes the rumen intact; aportion is metabolized to propionate (7).Propylene glycol escaping rumen fermentationis converted to glucose by the liver, primarilyvia the lactaldehyde pathway and subsequentoxidation to lactate (18). Administration of PGto postpartum cows reduced plasma concentra-tions of NEFA and ketones (27). Addition ofPG to the ration, compared with dietary starch,produced a greater molar percentage of rumenpropionate and elicited a greater insulin re-sponse (9). The objective of our study was todetermine whether prepartum administration ofPG could reduce periparturient fatty liver.

MATERIALS AND METHODS

Experimental De81gn end Sampling Protocol

Twenty-four multiparous Holstein cows,pregnant and nonlactating, were initially pairedaccording to calving date and assigned eitherto a control (n =11) or PG treatment (n =13)group. However, one cow assigned to the con-trol group was mistakenly treated, leaving anunbr!lanced design. Cows began the trial at 21d p:i.or to expected calving and were fedalfaL"a-bromegrass hay for ad libitum intakeand 1.8 kgld of concentrate separately to meetNRC (20) dry cow requirements. The prepar-tum concentrate contained 34% soybean meal,18.3% cracked corn, 18.3% soybean hulls,8.7% corn gluten meal, 8.7% blood meal, 8.7%meat and bone meal, 1.8% trace-mineralizedsalt, 1.3% vitamins A, D, and E (2,643,171 IUIkg of vitamin A, 881,057 IU/kg of vitamin D,and 881 IU/kg of vitamin E), and .3% mag-nesium oxide (OM basis). Following calving,cows were fed 50% alfalfa silage and 50%concentrate (OM basis) as a TMR for ad libi-tum intake. The postpartum concentrate con-tained 77.4% shell corn, 7.0% molasses, 6.0%

Journal of Dairy Science Vol. 76. No. 10. 1993

liquid protein supplement, 4.7% soybean meal,2.3% tallow, 1.6% sodium phosphate, .5%magnesium oxide, and .5% urea (OM basis).Feed and orts from each cow were sampledthree times weekly. Three weekly sampleswere composited into one sample for eachweek. Forage and concentrate DM were deter-mined by oven-drying at 60C for calculationof DM!. Cows received a l-L oral drench ofPG (Phoenix Pharmaceutical, Inc., S1. Joseph,MO) or water once daily at 1230 h from 7 dprior to expected calving until parturition.

Blood and plasma samples were asdescribed by Skaar et al. (29) and were takenvia the coccygeal vein on d 16 and 17 prior toexpected calving. Blood was also sampleddaily from 7 d prepartum to 3 d postpartumand on d 7, 14, and 21 postpartum. Duringtreatment, blood samples were taken 90 minpostdrenching. On d 5 prior to expected calv-ing, additional blood was sampled at IS-minintervals postdrenching for 2.5 h. One cowassigned to the control group calved beforemultiple sampling. Liver was sampled bypuncture biopsy at 17 d prior to expectedcalving and at 1 and 21 d postpartum. Allplasma samples were analyzed in random or-der. Plasma was analyzed for glucose andNEFA (29). Plasma was deproteinized (29) fol-lowing storage at -20C and then immediatelyanalyzed for BHBA (11). Plasma insulin wasanalyzed by radioimmunoassay (RIA) with amodification to the kit procedure (Coat-a-Count; Diagnostic Products Co., Los An-geles, CA) to account for relatively low insulinconcentrations in bovine plasma. Zero calibra-tor (100 JoLI) was added to 200 JoLI of eachstandard; the standard concentration wasdiluted by a factor of .66. Plasma (300 JoLI) wasthen used in the determination of each un-known sample. Insulin RIA intraassay varia-tion was 7.9%, and interassay variation was6.8%. Glucagon was analyzed by RIA accord-ing to the method of Reynolds et al. (25),except that Unger 30 K antibody was used.Glucagon RIA intraassay variation was 8.2%,and interassay variation was 6.0%. Liver tissuewas assayed for total lipid and TG (29).

Milk production was recorded daily through21 d postpartum. Milk samples were taken ata.m. and p.m. milkings on d 13, 14,20, and 21postpartum and analyzed for fat and proteinpercentages and somatic cells by infrared spec-trophotometry (Wisconsin DHI Cooperative

EFFECT OF PROPYLENE GLYCOL ON FATTY LIVER 2933

Laboratory, Appleton, WI). Body weight andbody condition score (6), averaged from twoevaluators, were recorded once prior to treat-ment.

Statistical Analysis

Total liver lipid and TG were analyzed in-dependently within sample period, i.e., 1 or 21d postpartum, according to the followingmodel: Yi = p. + b + Ti + ej where Yi = thedependent variable, p. =the overall mean of thepopulation, b = the regression coefficient forthe liver variable measured at d 17 prior toexpected calving, Ti = the average effect oftreatment i, and c; = the unexplained residualelement assumed to be independent and nor-mally distributed.

The OMI, analyzed independently withinsample period (i.e., pre- or postpartum), milkproduction, and milk composition were ana-lyzed according to the following model: Yijk =p. + Ti + Cj + Pic + TPik. + ejjlc where Yijlc =thedependent variable, p. =the overall mean of thepopulation, Ti =the average effect of treatmenti, Cj = the average effect of cow j, Pic = theaverage effect of period k, and c;jlc = theunexplained residual element assumed to beindependent and normally distributed. Analysisof covariance using fIrst lactation 305-d ma-ture equivalent milk production as the regres-sion coeffIcient was not successful in demon-strating differences in milk production. Theregression coeffIcient was not included in thefInal statistical model for milk production.

Period represented day for OMI and milkproduction and week for milk composition andFCM.

Plasma metabolites and hormones were ana-lyzed independently within sample period (i.e.,pre- or postpartum) according to the followingmodel: Yijk =p. + b + Ti + Cj + Pk + TPik +c;jk where Yijk =the dependent variable, p. =the overall mean of the population, b = theregression coefficient for the plasma variableaveraged between measurements on d 16 and17 prior to expected calving for cow j, Ti =theaverage effect of treatment i, Cj = the averageeffect of cow j, Pic = the average effect ofperiod k, and eijk = the unexplained residualelement assumed to be independent and nor-mally distributed. Period represented day fordaily plasma measurements and 15-min inter-val for multiple plasma samples. Models wereoriginally constructed to contain pair (i.e.,block), but the term was dropped because ofevents resulting in an uneven experimental de-sign. Variation between cows was used as theerror term when treatment effects on OMI,milk production and composition, and allplasma and liver measurements were tested.Differences were significant at P < .05, indi-cated a tendency at P = .05 to .15, and werenonsignifIcant at P > .15.

RESULTS AND DISCUSSION

Average BW and body condition score,measured prior to treatment, were not differentbetween treatment groups (Table I). Cows in

TABLE I. Description of experimental cows.

ControJi PG PX SD X SD

BW (pretreatment). kg 712 83 699 64 .69Body condition2 (pretreatment) 3.2 .6 3.1 .5 .68Parity 4.1 1.3 3.8 .8 .47First lactation perfonnance

7349 917 .05305-d Milk, kg 8045 745305-d 3.5% FCM. Icg 8230 819 7724 915 .17305-d ME MiIk.3 Icg 10,086 922 9663 1186 .35Fat. % 3.6 .2 3.8 .3 .02Protein. % 3.1 .I 3.2 .I .32

lOne liter of water (conlrol) or propylene glycol (FG) given once daily as an oral drench beginning at an average of 8d (conlrol) or 10 d (FG) prepartum through parturition.

2Scale ranged from 1 = thin to 5 = obese.3Mature equivalent.

Joumal of Dairy Science Vol. 76. No. 10. 1993

2934 STUDER ET AL.

5+-....,.-..,......,.-,......,.-..,.-...,.-..,.-..,.--.-25 -20 -15 -10 -5 0 5 10 15 20 25

Day relative to parturition

9'6C.Eai 8..0U::lC. 7..

E....

ii:6

Figure 2. Plasma glucose concentration of cows givena drench of 1 L of water (0) or propylene i!ycol (e) oncedaily begiMing at an average of 8 2.5 d (X SO; range,2 to II d) prepartum for cows assigned to the controlgroup and 10 3.6 d (range, 5 to 15 d) prepartum for cowstreated with propylene glycol through parturition (unad-justed means). A prepartum treatment effect (P < .01)occurred. Postpartum plasma glucose concentrations werenot significantly different between groups. The pre- andpostpartum standard errors were 2.1 and 1.9 mg/dl. respec-tively.

cally in both groups at parturition (Figure 2),which agrees with results of Schwalm andSchultz (28). Increased plasma glucose concen-tration could be a result of increased gluconeo-genesis, glycogenolysis, or both, stimulated bycatecholamines and glucocorticoids (21) at par-turition.

Plasma insulin concentrations were signifi-cantly elevated by 15 min following PG ad-ministration (Figure 4). Insulin concentrationpeaked prior to peak glucose concentration.Significant amounts of PG may have beenmetabolized to propionate, which stimulatespancreatic insulin secretion (14), or PG or in-termediates of PG metabolism might be directinsulin secretagogues. Insulin concentrationswere depressed on the day of calving (Figure5), in contrast to results of Kunz et al. (17),who reported an insulin surge at calving. Ourblood samples were taken at 24-h intervals;therefore, an acute insulin increase might nothave been detected. Plasma insulin concentra-tions were not different between treatmentgroups following parturition.

Prepartum plasma glucagon concentrationwas not affected by prepartum PG administra-tion (Figure 6), but daily samples were ob-

5D+....,.....-r-....,-.,........,.-..,..-r-..,.-..,.--.-25 -20 -15 -10 -5 0 5 10 15 20 25

Day relative to parturition

2

2

Figure 1. The OM! of cows given a drench of 1 L ofwater (0) or propylene glycol (e) once daily beginning atan average of 8 2.5 d ex SO; range, 2 to II d)prepartum for cows assigned to the control group and 10 3.6 d (range, 5 to 15 d) prepartum for cows treated withpropylene glycol through parturition (unadjusted means).Differences in DMI were not significant between groupspre- or postpartum. The pre- and postpartum standarderrors were .5 and .9 kg/d, respectively.

the control group produced significantly more305-d milk, and cows treated with PG had agreater fat percentage in their first lactation(Table 1), but 305-d 3.5% FCM and matureequivalent milk production were not differentbetween groups. Treatment duration was 8 2.5 d (X SO; range = 2 to 11 d) for thecontrol group and 10 3.6 d (range =5 to 15d) for the PG group. Multiple blood sampleswere obtained on d 6 1.5 (range = 3 to 8)prepartum for the control group and d 7 3.8(range = 1 to 12) prepartum for the PG group.

Fisher et al. (8) reported a depression inforage OMI when PG was fed at 9% of theconcentrate mixture (460 mlId). However, tominimize the potential for confounding effectsof PG and OMI, we delivered PG as an oraldrench. Both groups experienced a moderatedepression in feed intake immediately prepar-tum, but the depression was similar betweentreatments (Figure I).

Plasma glucose was elevated by PGthroughout treatment (Figure 2). Plasma glu-cose concentrations peaked at 75 min post-drenching (Figure 3). The elevation of plasmaglucose agrees with that of other reports (9,27), indicating that PG is an effective gluco-genic agent. Plasma glucose increased dramati-

:iiic

Journal of Dairy Science Vol. 76, No. 10, 1993

EFFECT OF PROPYLENE GLYCOL ON FATTY LIVER 2935

.80

.70

EQ.60

c:

.5'3

.50WI.5..E

.40WI..

a::.30

.20-25 20 15 105 0 5 10 15 20 25

Day relative to partu rillon

Figure 5. Plasma insulin concentration of cows given adrench of 1 L of water (0) or propylene &!ycol (e) oncedaily beginning at an average of 8 2.5 d (X SO; range.2 to 11 d) prepartum for cows assigned to the controlgroup and 10 3.6 d (range, 5 to 15 d) prepartum for cowstreated with propylene glycol through parturition(un-adjusted means). A prepartum treatment effect (P < .001)occurred. Postpartum plasma insulin concentrations werenot significantly different between groups. The pre- andpostpartum standard errors were .034 and .030 ng/mI.respectively.

8

7;;QEoi 7..0U::I'6>.. 6E....

a::6

5s+-......r--"'T""--r-T"""....,..-r-""T'"--.r--"'T"""""o 15 30 45 60 75 90 105120 135150

Postdrenching, min

Figure 3. Plasma glucose concentration of cows givena dren

2936 STUDER ET AL.

-25 20 -15 -10 -5 0 5 10 15 20 25Day Relative 10 Parturition

Figure 8. Plasma NEFA concentration of cows given adrench of 1 L of water (0) or propylene &!ycol (e) oncedaily beginning at an average of 8 2.5 d (X SO; range.2 to 11 d) prepartum for cows assigned to the controlgroup and 10 3.6 d (range. 5 to 15 d) prepartum for cowstreated with propylene glycol through parturition (unad-justed means). A prepartum treatment effect (f' < .001). apostpartum treatment effect (f' < .10), and a postpartumtreatment by time interaction (P < .10) occurred. The pre-and postpartum standard errors were 28.8 and 52.7 pM,respectively.

tained at 24-h intervals, and acute changes inplasma glucagon might not have been detected.However, differences between groups inplasma glucagon concentration were not de-tected in the prepartum multiple sample period(Figure 7). Plasma glucagon concentrations(Figure 6) were elevated as milk productionand OMI increased, and a significant day ef-fect existed across pre- and postpartum peri-ods. Increasing blood VFA concentrations actdirectly on the pancreas to stimulate glucagonsecretion in ruminants (14). Glucagon concen-trations increase postpartum, which accompa-nies decreased plasma glucose and insulin.Glucose inhibits glucagon secretion (22).

Plasma NEFA concentration elevatedgradually between d 16 and 2 prepartum incows in the control group, but not in cowstreated with PG (Figure 8). As previouslynoted by Reid et al. (24), the increase inplasma NEFA was sharp at calving. The in-crease in NEFA concentration between d 16and 2 prepartum may be attributed to an in-crease in fetal glucose demand, followed bysubsequent maternal reliance on fatty acids(30), depression in feed intake, possiblecatacholamine-stimulated lipolysis associatedwith stress at calving, or a combination ofthese factors. The relative contribution of the

80

::I!"-

.(u. 60wz..

E.... 40ii:

O+-...,.-,.-..,......,-.,-..,.-,.-............ro--..-25 -20 -15 -10 -5 0 5 10 15 20 25

Day relative 10 parturition

Figure 9. Plasma BHBA concentration of cows given adrench of I L of water (0) or propylene &!ycol (e) oncedaily beginning at an average of 8 2.5 d (X SO; range.2 to 11 d) prepartum for cows assigned to the controlgroup and 10 3.6 d (range. 5 to 15 d) prepartum for cowstreated with propylene glycol through parturition (unad-justed means). A prepartum treatment effect (P < .001)occurred. Postpartum plasma BHBA concentrations werenot significantly different between groups. The pre- andpostpartum standard errors were .7 and 2.1 mg/dl. respec-tively.

3

2

~'"

2E.(III 1:z:III..

E 1....

Ii:5

co

~::JQ..E:I;;:

.17 50101-""111""5--'3r"'"0--'4r"'"5"""60-""75-"'90-1"!0~5-1"'2~0-1~5"":"!150Posldrenchlng, min

Eg..22

.25

Figure 7. Plasma glucagon concentration of cowsgiven a drench of 1 L of water (0) or propylene glycol (e)on d 6 1.5 (X SD; range. 3 to 8) prepartum for cowsassigned to the control group and d 7 3.8 (range. 1 to 12)prepartum for cows treated with propylene glycol (unad-justed means). Plasma glucagon concentrations were notsignificantly different between groups. The overall stan-dard error was .009 ng/ml.

Journal of Oairy Science Vol. 76. No. 10, 1993

EFFECT OF PROPYLENE GLYCOL ON FAITY LNER 2937

TABLE 2. Effect of prepartum propylene glycol (PG) administration on total liver lipid and liver triglyceride (TO)concentration (DM basis).l

ControI2 SE

Total liver lipid. %Day relative to parturition-16 (covariate period) 15.9 .4+01 29.5 2.3+21 31.4 2.8

Liver TG. %-16 (covariate period) 1.6 .2+01 14.5 2.5+21 17.3 2.2

21.122.3

5.07.2

SE

2.12.1

2.32.1

p

.01

.01

.01

.01

lDala for d -16 are pooled and were used for analysis of covariance. and data for d I and 21 are covariately adjustedmeans.

2()ne liter of water (control) or PO given once daily as an oral drench beginning at an average of 8 d (control) or 10 d(PG) prepartum through parturition.

gradual increase in plasma NEFA concentra-tion prior to parturition, or the acute surge atparturition, to development of hepatic TG ac-cumulation is not known. Plasma NEFA con-centration of cows treated with PG was signifi-cantly lower than that of cows in the controlgroup prior to parturition, most likely becauseof inhibition of adipose adenylate cyclase ac-tivity and lipolysis by elevated plasma insulin(32). We have no explanation for the tendencytoward lower postpartum plasma NEFA con-centration in cows receiving PG because treat-ments were terminated at calving.

Prepartum plasma BHBA concentrationswere decreased by PG treatment (Figure 9).Lower BHBA concentrations may reflecthigher concentrations of insulin and lower con-centrations of NEFA. Insulin may act to reduceNEFA supply to the liver, to inhibit hepaticketogenesis, to increase ketone utilization byperipheral tissue (4), or trigger a combinationof these effects. Postpartum plasma BHBAconcentrations were similar between treat-ments.

At 1 d postpartum, total liver lipid and TGwere significantly lower in cows treated withPG, and the differences were maintainedthrough 21 d postpartum (Table 2). The differ-ence in liver TG content postpartum accompa-nied a difference in plasma NEFA concentra-tion between treatment groups. Changes inliver TG are a more sensitive measure ofhepatic lipid accretion because they ignorephospholipid and cholesterol, which do not

change markedly during elevated NEFA up-take (3). Both groups experienced a smallerincrease in liver TG concentration between 1and 21 d postpartum, which supports the con-cept that the majority of hepatic TG depositionoccurs between 16 d prepartum and 1 d post-partum (12).

Correlation analysis was performed to iden-tify prepartum parameters related to peripar-turient hepatic TG accumulation (Table 3).Prepartum plasma NEFA concentration washighly correlated with liver TG at 1 d postpar-tum. The likely importance of maintaining

TABLE 3. Correlation coefficients between prepartumplasma metabolites. hormones. DMI. or body conditionscore and liver triglyceride (TO) concentrations at I and 21d postpartum.

Liver TG

Prepartum 1 d 21 dparameter1 postpartum postpartum

r p r PGlucose -.49 .02 -.36 .09NEFA .46 .03 .11 .60BHBA .36 .08 .45 .03Insulin -.49 .02 -.09 .69Glucagon .03 .91 -.07 .75DMI -.25 .24 -.22 .31Body condition

score2 .23 .29 .12 .56

lplasrna concentrations and DMI are expressed asmeans between d I and 7 prepartum.

2Scale ranged from 1 = thin to 5 = obese.

Journal of Dairy Science Vol. 76. No. 10. 1993

2938 STUDER ET AL.

TABLE 4. Postpartum milk production. milk composition. and sec.

Control I SE pol SE P

Milk production. kgld 33.2 2.1 32.6 1.8 .753.5% FCM. kgld 38.7 1.4 38.0 1.6 .69Fat, % 3.97 .19 3.86 .11 .59Fat. kgld 1.42 .06 1.38 .06 .66Protein. % 3.10 .10 3.11 .08 .86Protein. kgld 1.12 .03 1.11 .03 .95SCC. lOOO1m1 396 181 426 194 .90

lOne liter of water (control) or propylene glycol (PG) given once daily as an oral drench beginning at an average of 8d (control) or 10 d (PG) prepartum through parturition.

prepartum plasma glucose and insulin concen-trations in indicated by their strong negativecorrelation with liver TG at 1 d postpartum,but not at 21 d postpartum. Plasma BHBAconcentration and liver TG accumulation werepositively correlated at 1 and 21 d postpartum.

Milk production and composition were notaffected by prepartum PG administration (Ta-ble 4). In a previous study (2), cows force-fedvia rumen cannulas had a lower liver TG con-tent at 1 d postpartum and produced more milkthrough 28 d postpartum. Longer postpartumperiods would be desirable for detection oftreatment effects on milk production. Cowsused in this experiment were in moderate bodycondition (Table 1). Production benefits fromPG administration may have been realized ifexperimental cows had been in excessive bodycondition, possibly predisposing them to fattyliver and ketosis (24). Differences betweengroups in animal health were not observed.

The PG dose given in this study was inexcess of the standard postpartum field recom-mendation of 230 to 410 mUd (15). The mini-mum effective prepartum dose of PG for thealleviation of fatty liver is unknown. A sum-mary of data from our laboratory and otherreports (9, 26, 27) indicate that the antilipolyticeffect of PG increases linearly from 117 mUdof PG to 1 Ud of PG (r2 = .86).

CONCLUSIONS

Prepartum PG administration reducedhepatic TG accumulation by 32 and 42% at 1and 21 d postpartum, respectively. Evidencefrom this experiment suggests that reducingprepartum plasma NEFA concentration is im-portant for reducing periparturient hepatic TG

Journal of Dairy Science Vol. 76. No. 10. 1993

accumulation. This reduction could beachieved by improved energy balance in lategestation through maintenance of prepartumDMI, enhanced carbohydrate and insulin statusthrough administration of glucogenic precur-sors during the last few days of gestation, orboth. Increasing dietary nonstructural carbohy-drates may have limited effectiveness, con-sidering the inherent decline in feed intakeimmediately prepartum and complications thatcan arise from rumen acidosis. Administrationof PG elevated prepartum plasma glucose con-centrations; however, the stimulation of insulinsecretion probably more directly modulatedprepartum plasma NEFA concentration. Fur-ther research is needed to evaluate the possiblerole of insulin as a preventative agent in thealleviation of periparturient fatty liver.

ACKNOWLEDGMENTS

The care and feeding of the experimentalcows was coordinated by Leland Danz, whosecontribution is greatly appreciated. The authorsalso thank Dawn Miksic for her assistance inthe laboratory.

REFERENCES

I Bell.A. W. 1980. Lipid metabolism in the liver andselected tissues and in the whole body of ruminantanimals. Prog. Lipid Res. 18:117.

2 Bertics. S. J. R. R. Grummer. C. Cadomiga-Valino.and E. E. Stoddard. 1992. Effect of prepartum drymatter intake on liver triglyceride concentration andearly lactation. 1. Dairy Sci. 75:1914.

3 Butler. S. M. A. Faulkner. V. A. Zammit, and R. G.Vernon. 1988. Fatty acid metabolism of the perfusedcaudate lobe from livers of fed and fasted non-pregnant and fasted late pregnant ewes. Compo Bio-chern. Pbysiol. B. Compo Biochem. 91:25.

EPFECT OF PROPYLENE GLYCOL ON FATIY LIVER 29394 Brockman, R. P. 1979. Roles of insulin and glucagon

in the development of ruminant ketosis: a review.Can. Vet. 1. 20:121.

5 Cadomiga, C. P., R. R. Gnunmer, L. E. Armentano,and S. S. Donkin. 1992. Effects of extracellular fattyacid concentration, estriol, insulin, and glucagon onhepatic triglyceride accumulation, acid soluble prod-uct fonnation, and gluconeogenesis. 1. Dairy Sci.75(Suppl. 1):183.(Abstr.)

6 Edmonson, A. 1., I. 1. Lean, L. D. Weaver, T. Farver,and G. Webster. 1989. A body condition scoring chartfor Holstein dairy cows. J. Dairy Sci. 72:68.

7 Emery, R. S., N. Burg, L. D. Brown, and G. N. Blank.1964. Detection, occurrence, and prophylactic treat-ment of borderline ketosis with propylene glycol feed-ing. J. Dairy Sci. 47:1074.

8 Fisher, L. 1., 1. D. Ertle, G. A. Lodge, and F. D.Sauer. 1973. Effects of propylene glycol or glycerolsupplementation of the diet of dairy cows on feedintake, milk yield and composition, and incidence ofketosis. Can. 1. Anim. Sci. 53:289.

9 Fronk, T. 1. 1975. The effect of propylene glycol onmilk fat depression. M.S. Thesis, Univ. Wisconsin,Madison.

10 Gerloff, B. 1., T. H. Herdt, and R. S. Emery. 1986.The relationship of hepatic lipidosis to health andperformance in dairy cattle. 1. Am. Vet. Med. Assoc.188:845.

11 Gibbard, S., and P. 1. Watkins. 1968. A micro-methodfor the enzymatic determination of d-fJ-hydroxybutyrale and acetoacetate. Oin. Chirn. Acta19:511.

12 Grummer, R. R. 1993. Etiology of lipid-related meta-bolic disorders in periparturient cattle. J. Dairy Sci. 76(Nov).

13 Grummer, R. R., S. 1. Bertics, D. W. LaCount, I.A.Snow, M. R. Dentine, and R. H. Stauffacher. 1990.Estrogen induction of fatty liver in dairy cattle. 1.Dairy Sci. 73:1537.

14 Harmon, D. L. 1992. Impact of nutrition on pancreaticexocrine and endocrine secretion in ruminants: a re-view. 1. Anim. Sci. 70:1290.

15 Herdt, T. H., and R. S. Emery. 1992. Therapy ofdiseases of ruminant intennediary metabolism. Vet.Clin. North Am. Food Anim. Pract. 8:91.

16 Kleppe, B. B., R. 1. Aiello, R. R. Grummer, and L. E.Armentano. 1988. Triglyceride accumulation and verylow density lipoprotein secretion by rat and goathepatocytes in vitro. 1. Dairy Sci. 71:1813.

17 Kunz, P. L., 1. W. Blum, I. C. Hart, H. Bickel, and J.Landis. 1985. Effects of different energy intakes be-fore and after calving on food intake, performance andblood hormones and metabolites in dairy cows. Anim.Prod. 40:219.

18 Miller, O. N., and G. Bazzano. 1965. Propanediolmetabolism and its relation to lactic acid metabolism.Ann. New York Acad. Sci. 119:957.

19Morrow, D. A., D. Hillman, A. W. Dade, and H.Kitchen. 1979. Clinical investigation of a dairy herdwith the fat cow syndrome. J. Am. Vet. Med. Assoc.174:161.

20 National Research Council. 1989. Nutrient Require-ments of Dairy Cattle. 6th rev. ed. Nat!. Acad. Sci.,Washington, DC.

21 Norman, A. W., and G. Litwack. 1987. Adrenal corti-coids. Page 412 in Hormones. Academic Press, SanDiego, CA.

22 Pipeleers, D. G., F. C. Schuit, C. Van Schravendijk,and M. Van De Winkel. 1985. Interplay of nutrientsand hormones in the control of glucagon release.Endocrinology 117:817.

23 Reid, I. M, R. A. Collins, G. D. Baird, C. 1. Roberts,and H. W. Symonds. 1979. Lipid production rates andthe pathogenesis of fatty liver in fasted cows. 1. Agric.Sci. (Camb.) 93:253.

24 Reid, I. M., C. 1. Roberts, R. J. Treacher, and L. A.Williams. 1986. Effect of body condition at calving ontissue mobilization, development of fatty liver andblood chemistry of dairy cows. Anim. Prod. 43:7.

25 Reynolds, C. K., H. F. Tyrell, and L. E. Annentano.1992. Effects of mesenteric vein n-butyrate infusionon liver metabolism by beef steers. J. Anim. Sci. 70:2250.

26 Ruegsegger, G. 1., and L. H. Schultz. 1986. Use of acombination of propylene glycol and niacin for sub-clinical ketosis. 1. Dairy Sci. 69: 1411.

27 Sauer, F. D., 1. D. Ertle, and L. 1. Fisher. 1973.Propylene glycol and glycerol as a feed additive forlactating dairy cows: an evaluation of blood metabo-lite parameters. Can. 1. Anim. Sci. 53:265.

28 Schwalm, 1. W., and L. H. Schultz. 1976. Relationshipof insulin concentration to blood metabolites in thedairy cow. J. Dairy Sci. 59:255.

29 Skaar, T. C., R. R. Grummer, M. R. Dentine, and R.H. Stauffacher. 1989. seasonal effects of prepartumand postpartum fat and niacin feeding on lactationperformance and lipid metabolism. 1. Dairy Sci. 72:2028.

30 Speroff, L. 1981. The ovary. Page 684 in Endocrinol-ogy and Metabolism. P. Felig, 1. D. Baxter, A. E.Broadus, and L. A. Frohman, ed. McGraw-Hili, NewYork, NY.

31 Veenhuizen, 1. 1., J. K. Dracldey, M. J. Richard, T. P.Sanderson, L. D. Miller, and 1. W. Young. 1991.Metabolic changes in blood and liver during develop-ment and the treatment of experimental fatty liver andketosis in cows. 1. Dairy Sci. 74:4238.

32 Vernon, R. G. 1980. Lipid metabolism in the adiposetissue of ruminant animals. Prog. Lipid Res. 19:23.

10urnal of Dairy Science Vol. 76, No. 10, 1993