METABOLIC DISORDERS OF RUMINANTS 0749--0720/00 $15.00 + .

00

SUBCLINICAL KETOSIS IN
LACTATING DAIRY CATTLE
Todd Duffield, DVM, DVSc

Subclinical ketosis is simply a condition marked by increased levels

of circulating ketone bodies without the presence of the clinical signs of
ketosis. Subclinical ketosis causes economic losses through decreased
milk production and association with periparturient diseases. It has a
reported worldwide prevalence of 8.9% to 34% for cows in the first 2
months of lactation,5, 20, 44 whereas the reported lactational incidence of
clinical ketosis varies from 2% to 15%.8
As the dairy industry continues to strive for increased genetic gains
in milk production, cows will continue to be at increased risk of devel-
oping ketosis. Uribe et al examined data from over 7000 Canadian
Holsteins and found that there was a genetic antagonism between the
production of milk and milk fat and clinical ketosis. 8o Ketosis did not
have a high heritability (h2 = 0.09), but the risk of ketosis increased with
increases in milk production and milk fat yield. 80 The authors concluded
that long-term increases in ketosis incidence can be expected if genetic
selection is based solely on increasing milk and milk fat production
traits. 80 Prevention strategies that will reduce the incidence of subclinical
and clinical ketosis while maintaining high levels of milk production
should be beneficial to the dairy industry.

CAUSE OF KETOSIS

Elevated levels of circulating ketone bodies occur in early lactation

in response to the homeorhetic drive to sustain high levels of milk

From the Department of Population Medicine, Ontario Veterinary College, University of

Guelph, Guelph, Ontario, Canada

VETERINARY CLINICS OF NORTH AMERICA:

FOOD ANIMAL PRACTICE

VOLUME 16 • NUMBER 2 • JULY 2000 231

232 DUFFIELD

production at a time when dry matter intake is reduced. 8 The three

major ketone bodies are acetone, acetoacetate, and ~-hydroxybutyrate
(BHB). Ketone bodies represent an integral part of ruminant intermedi-
ary metabolism and they provide an important form of energy to periph-
eral tissues when carbohydrate levels are reduced. In addition to contrib-
uting to tissue energy supply, BHB is used for milk fat synthesis,60 so a
baseline level of circulating ketone bodies is normal in the ruminant. See
the article by Herdt (pp 215-230) elsewhere in this issue for a more
detailed discussion of the cause of ketosis.

DEFINITION OF SUBCLINICAL KETOSIS

Subclinical ketosis in dairy cattle is characterized by the presence of

an excess level of circulating ketone bodies in the absence of the clinical
signs of ketosis. 4 Lactating cows with clinical ketosis typically display
diminished appetite, especially for concentrates, frequently have hard
dry feces, and show both a noticeable decrease in milk yield and a
marked loss of body weight. l l Occasionally, cows with clinical ketosis
will also display nervous signs, such as vigorous licking, and apparent
blindnessY It is not simply the presence but an abnormal level of
circulating ketone bodies that determines subclinical ketosis. When ke-
tone bodies are measured quantitatively, a defined threshold must be
used to separate normal animals from those with subclinical ketosis.
Both clinical ketosis and subclinical ketosis have been previously re-
viewed.4, 8, 51, 52

MEASUREMENT OF KETONE BODIES

Ketone bodies are present in blood, urine, and milk. 12 A range of

thresholds and methods for detection of subclinically elevated ketone
body concentrations have been reported in the literature. BHB is the
predominant circulating blood ketone body, but there is a strong correla-
tion between the whole blood concentrations of BHB and acetoacetate44;
however, acetoacetate increases more compared with BHB with increas-
ing levels of total ketone bodies, despite a continuing rise in BHB.6 This
is most likely because of a relative increase in liver ketogenesis through
increased fatty acid metabolism as well as a reduction or cessation in
BHB synthesis in the ruminal epithelium. 7 Acetoacetate is unstable in
both tissues and samples. It readily decomposes to acetone and carbon
dioxide. 10 By contrast, BHB is relatively stable in whole blood, plasma,
or serum, both in vivo and in vitro. 17, 79 BHB measured using an enzy-
matic method in either serum or whole blood provides similar results;
however, BHB measured with the same test in plasma is generally
systematically lower, possibly because of interference from anticoagu-
lants. 17 The presence of hemolysis in serum has been shown to interfere
with BHB analysis, causing elevated measurements. 24 All ketone bodies
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 233

show some diurnal variation in relation to feed intake, but BHB demon-
strates the most marked diurnal variations of all the ketone bodies.
Peak levels of BHB occur roughly 4 hours post feeding. 79, 81 Andersson
determined that milk acetone was approximately 95% of blood acetone,
milk acetoacetate was 45% of blood acetoacetate, and blood acetoacetate
was 13% of blood BHB.3 BHB levels in milk are only 100/0 to 15% of
circulating blood levels, possibly because of the ketone body's role in fat
metabolism. 3 There is a high coefficient of correlation between circulating
levels of acetone plus acetoacetate and BHB.3,68

DIAGNOSING SUBCLINICAL KETOSIS AND CRITICAL

SUBCLINICAL THRESHOLDS

Serum Il-Hydroxybutyrate

Based on several studies, subclinical ketosis may start at serum BHB

concentrations above 1000 J..Lmol/L (10.4 mg/ dL) and clinical ketosis at
about 2600 J..Lmol/L (27 mg/ dL); however, at exactly what level individ-
ual cows will express clinical signs is extremely variable. 3 Furthermore,
the values used to set an appropriate subclinical threshold using serum
or plasma BHB appear to be somewhat arbitrary. Studies using BHB for
assessing subclinical ketosis report a range of values from 1000 J..Lmol/L
(10.4 mg/ dL)47, 81 to 1400 J..Lmol/L (15 mg/ dL)82 for defining a subclinical
threshold. Kelly suggested that 1000 J..Lmol/L (10.4 mg/ dL) be used to
separate cows with low and high BHB concentrations.47 Whitaker et al
chose 1000 J..Lmol/L BHB (10.4 mg/ dL) as a cutoff point to identify
subclinical ketosis in groups of early lactation cows based on anecdotal
experience. 81 Plasma BHB concentrations above 1400 J..Lmol/L (15 mg/
dL) combined with plasma glucose values of less than 3.0 mmol/ L
(55mg/ dL) were used to classify cows as having poor energy status. 82
Nielen et al selected a subclinical ketosis threshold of 1200 J..Lmol/ L (12.5
mg / dL) BHB based on the distribution of their data. 58 Less arbitrary
threshold values of BHB for defining subclinical ketosis have been deter-
mined recently. In a 1010-cow study, a serum concentration of 1400
J..Lmol/L BHB (15 mg/ dL) or greater in the first 2 weeks postcalving was
found to cause a threefold greater risk for cows to subsequently develop
either clinical ketosis or abomasal displacement. 23 In addition, cows
having serum BHB concentrations at or above 2000 J..Lmol / L (21 mg / dL)
within 7 days prior to the first Dairy Herd Improvement (DHI) test
produced over 4 kg less milk on DHI test day.

Blood Acetoacetate

Despite its instability, blood acetoacetate levels have been used by

some authors to identify animals with subclinical ketosis. Baird has
suggested a threshold of 500 J..Lmol/L (5.0 mg/ dL) blood acetoacetate
234 DUFFIELD

for hyperketonemic cows to display clinical signs of ketosis. 8 Kauppinen

chose 360 J.Lmol/L (3.7 mg/ dL) and 1050 J.Lmol/L (10.1 mg/ dL) blood
acetoacetate to define lower thresholds for subclinical and clinical keto-
sis, respectively.45 At the ratios of acetoacetate to BHB commonly ob-
served in jugular vein blood, these thresholds would approximate 2800
J.Lmol/L (29 mg/ dL) BHB for subclinical ketosis and 8000 J.Lmol/L (83
mg/ dL) BHB for clinical ketosis. These BHB values are very high and
may be the result of using mammary (subcutaneous abdominal) vein
samples. It has been shown that the mammary gland tends to extract
BHB but releases acetoacetate,50 which would lead to a higher BHB-to-
acetoacetate ratio in mammary vein blood than in jugular blood.

Milk Ketone Body Tests

Elevated ketone body concentrations can also be detected with milk

tests that use the reaction of acetone and acetoacetate with sodium
nitroprusside. 2,68 Studies using this test for detection of subclinical keto-
sis have a threshold value automatically determined by the detection
limit of the test itself; however, the test can be interpreted in a semiquan-
titative manner because higher concentrations of milk ketone bodies
have been reported with increased intensity of the nitroprusside reac-
tion. 2 Nitroprusside reacts more with acetoacetate than with acetone2
and is completely nonreactive with BHB.12 The nitroprusside test can
detect 500 J.Lmol/ L (4 mg / dL) of total milk ketone bodies, and 300
J.Lmol/ L (2 mg / dL) of milk acetoacetate and acetone, which represents
1000 to 1500 J.Lmol/L (10 to 15 mg/ dL) total blood ketone bodies. 68 Both
Schultz and Myers68 and Andersson3 reported high correlations between
blood ketone bodies and milk acetone plus acetoacetate. This implies
that measurement of milk acetone and acetoacetate should reflect circu-
lating levels of blood ketone bodies. There is a new, rapid milk test that
measures BHB levels in milk semiquantitatively.19 The test detects milk
BHB values at 50, 100, 200, 500, and 1000 J.Lmol/L (0.5, 1.0, 2.0, 5.0, and
10.0 mg/ dL).

Automated Milk Acetone

Marstorp et al developed an injection flow spectrophotometric sys-

tem to measure milk acetone. 55 This automated analysis has been a
popular methodology in Scandinavian ketosis research. Andersson and
Emanuelson used a threshold of 400 J.Lmol/ L (4 mg / dL) of milk acetone
to distinguish normal from hyperketonemic cows. 5 This level was also
used in two studies by Gustafsson et al; however, the researchers sug-
gested that 400 J.Lmol / L (4 mg / dL) might be slightly low, because
minimal production losses were experienced in the acetone class of 400
to 1000 J.Lmol/L (4 to 10 mg/ dL) compared with the higher losses in
milk yield above 1000 J.Lmol/L (10 mg/ dL) milk acetone. 40,41 Gustafsson
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 235

determined that the critical borderline concentration for acetone is 700

/-Lmol/L (0.7 mg/ dL) based on first and second production tests in early
lactation. 39 Steen et al also used 700 /-Lmol/L (0.7 mg/ dL) milk acetone
to separate normal from hyperketonemic COWS.72 Miettenen found a
continuous negative effect of milk acetone on milk yield but suggested
a threshold· for milk acetone of 50 /-Lmol/ L (0.3 mg / dL) for defining
subclinical ketosis. 56 This level was based on the expected milk yield of
only 10 cows, from which the lowest acetone level measured was 50
/-Lmol/ L. This suggested threshold appears unrealistically low. A milk
acetone threshold of 400 /-Lmol / L (2.4 mg / dL) would be the approximate
equivalent of 1500 /-Lmol/L (15 mg/ dL) serum BHB, and 700 /-Lmol/L of
milk acetone would approximate 2700 /-Lmol/L (28 mg/ dL) serum BHB.
These conversions of ketone body concentrations between type of ke-
tone, and between milk and blood, are based on estimated ratios that
have been previously reported. 3 The flow injection technology is not
readily available to dairy practitioners outside of Scandinavia. Current
research is being pursued for monitoring subclinical ketosis by measur-
ing milk acetone at regular DHI tests with the same machine that
measures milk components and milk urea nitrogen.

Other Tests for Subclinical Ketosis

A routine test for subclinical ketosis available through a central milk

testing facility would be a convenient method for monitoring energy
status of dairy herds. The Fossomatic 4000 milk analyzer (Foss Electric,
Hiller0d, Denmark) used by several milk testing facilities has the capa-
bility to measure milk citrate. Citrate plays an integral role in the energy
metabolism of the cell as a key component in the Krebs' cycle but the
biologic role of citrate in milk is essentially unknown. 32 Citrate is formed
from the joining of acetyl CoA with oxaloacetate.8 Theoretically, low
citrate levels would be present during ketosis because oxaloacetate is in
short supply; however, citrate has not been found to be closely associated
with subclinical ketosis and does not appear to useful for monitoring
subclinical ketosis. 23 Because milk fat and milk protein percentages are
altered in sublinical ketosis, these variables have been investigated for
their use in defining subclinical ketosis. Among all protein and fat
variables, a protein-to-fat ratio of 0.75 or greater was the best test for
diagnosing subclinical ketosis in a Canadian study22; however, the pro-
tein-to-fat ratio was not a good test overall, having a sensitivity of 58%
and a specificity of 69%.

Cowside Test Performance

The detection limits of the milk ketone body tests do not describe
their epidemiologic test characteristics. Nielen et al found that the nitro-
prusside milk test had a sensitivity of 90% and a specificity of 96% when
compared with serum BHB levels of 1400 /-Lmol/L (15 mg/ dL).58 This
236 DUFFIELD

excellent test performance was not repeated in subsequent Canadian

research. 23,35 Using Ketocheck (Great States, st. Joseph, MO) test results
compared with a serum BHB threshold of 1200 f.Lmol/ L (12 mg / dL) or
greater, Duffield reported that the sensitivity was only 380/0, whereas the
sensitivity at a BHB threshold of 2000 f.Lmol/L (21 mg/ dL) was 610/0. 23
The specificity was extremely high (99%). Geishauser et al compared
serum BHB to five cowside tests including the Utrecht formulation. 35
Although this formulation performed the best of the nitroprusside tests,
the sensitivity was only 42%. The best cowside test appears to be the
milk BHB test (Ketolac; Hoechst, UnterschleifSheim, Germany). Geis-
hauser et al report a range in sensitivity and specificity depending on
the cutoff level of Ketolac that is used. 35 Using 100 f.Lmol/L milk BHB
gives a sensitivity of 720/0 and a specificity of 89%. Increasing the milk
BHB cutoff improves the specificity at the expense of sensitivity, whereas
decreasing the milk BHB cutoff has the opposite effect. It has been
suggested that colostrum may interfere with the performance of the milk
ketone body test powder,68 and the sensitivity may not be as high when
used in the first week of lactation. Lower sensitivities were detected in
the first week postcalving using nitroprusside powder tests in a Guelph
study.23 A nitroprusside tablet for urine was 100% sensitive but only
59% specific in the study by Nielen et aL58 Urine ketone body concentra-
tions are four times higher than blood levels. 69 The high ketone body
concentrations and potential urinary metabolites that may react with
nitroprusside12 are likely reasons for the high percentage of false-positive
reactions. The poor specificity and relative inconvenience in sample
retrieval has limited the usefulness of urine ketone body evaluations in
field studies. Milk, by contrast, is relatively easy to obtain, and the
high correlation between milk and blood ketone bodies, as well as the
inexpensive cost of nitroprusside powder, has allowed these tests to gain
popularity; however, the currently available milk nitroprusside tests
suffer from low sensitivity that limits their usefulness for screening cows
in early lactation. A promising new liquid nitroprusside milk ketone
test (Pink Milk; www.profs-products.com) has shown both excellent
sensitivity and specificity (Dr. Thomas Geishauser, personal communica-
tion, 1999). Test performance characteristics of various cowside tests are
outlined in Table 1.

EPIDEMIOLOGY OF SUBCLINICAL KETOSIS

Prevalence and Incidence

Most large studies report the prevalence of hyperketonemia or hy-

perketolactia rather than subclinical ketosis. This is because the identity
of cows that displayed clinical signs is not generally known, and animals
with either clinical or subclinical ketosis are included. 4 Because there is
a high correlation between milk and blood ketone bodies, the terms
hyperketonemia and hyperketolactia can be considered interchangeable for
discussing the prevalence and incidence rates. The range in reported
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 237

Table 1. PERFORMANCE OF COWSIDE TESTS FOR IDENTIFYING SUBCLINICAL

KETOSIS

Subclinical Ketosis
Threshold (Serum Sensitivity Specificity Reference
Test BHB .... moI/L) (%) (%) No.
Utrecht nitroprusside-milk 1400 90 96 58
Utrecht nitroprusside-urine 1400 100 59 58
Bioketone 1200 33 100 35
Ketocheck 1200 28 100 35
Utrecht powder 1200 42 100 35
Ketostix 1200 5 100 35
Ketolac-50 J.Lmol/ L 1200 92 55 35
Ketolac-100 J.Lmol/L 1200 72 89 35
Ketolac-200 J.Lmol/L 1200 45 97 35
Ketolac-500 J.Lmol/ L 1200 17 100 35
Ketocheck 1400 38 99 23
Ketocheck 2000 61 98 23

Data from references 23, 35, and 58.

prevalence will vary according to the definition of hyperketonemia,

with lower thresholds yielding a higher prevalence. Hyperketonemia is
commonly reported for the first 2 months of lactation, because this is
the primary risk period. 8 Dohoo and Martin found more positive milk
ketone body tests in the first versus the second month of lactation and
observed that the peak prevalence of hyperketonemia occurred in the
third and fourth week of lactation. 20 Other researchers also found that
the peak prevalence of hyperketonemia occurred during the third and
fourth week postcalving5, 45, 70; however, recent work suggests that there
is an earlier peak that occurs in the first 2 weeks postcalving. 22,25 It may
be that advances in genetics and feeding management have pushed the
metabolic challenge closer to calving. The earlier peak occurrence of
hyperketonemia may also reflect a difference in the metabolic origin of
ketosis, as discussed in the article by Herdt (pp 215-230) elsewhere in
this issue. Reported prevalences for hyperketonemia in the first 2 months
of lactation range from 8.9% to 34% in various studies. 5, 20, 22, 45, 58 Dohoo
and Martin estimated that the minimum lactational incidence rate was
12.1%, and the maximum was 100%, depending on the duration of the
condition. 20 In that study, the interval between milk ketone tests was at
least 30 days for eight herds and approximately 90 days for the other 24
herds; a true lactational incidence of subclinical ketosis was difficult to
determine. Emery et al used weekly milk ketone testing and found a
lactational incidence of ketolactia of 29% in one herd. 29 Simensen et al
measured acetoacetate in weekly milk samples from cows within five
herds that had a high clinical ketosis incidence rate and discovered that
46% of cows in the first 6 weeks of lactation had milk acetoacetate values
above 100 fLmol/L.70 Emery et al suggested that 50% of all lactating
cows go through a stage of subclinical ketosis in early lactation.29 Cana-
dian research supports this notion. In 507 untreated Holstein cows and
heifers from 25 dairy herds, the cumulative incidence over the first
238 DUFFIELD

9 weeks of lactation was 590/0 and 43%, using cutoff threshold BHB
concentrations of 1200 and 1400 J.1mol/L, respectively.25 The peak inci-
dence occurred in the first week post calving (Fig. 1). The lactational
incidence and duration of subclinical ketosis will determine the preva-
lence of the condition. Dohoo and Martin estimated the duration to be
about 16 days, using a prevalence of 12.1 % and assuming the lactational
incidence to be 50%. 20

Risk Factors for Subclinical Ketosis

Important factors that influence the occurrence of subclinical ketosis

include herd, parity, body condition, genetics, and season. There are
many reports of herd differences in the prevalence of hyperketonemia. 5,
20, 41, 43, 72 Dohoo and Martin discovered a range in the herd prevalence of
ketolactia from 00/0 to 340/0.2° Cow and herd accounted for 14% and 6%
of the total variation in BHB in a Canadian study.23 Increasing the
feeding frequency and increased volume of concentrates fed have been
shown to reduce the incidence of hyperketonemia. 41 This could help
explain some differences in the prevalence and incidence rates of subclin-
ical ketosis among herds. Tveit et al reported an association between
year and level of acetoacetate and attributed this relationship to in-

35

30

---
~
c=
......., 25
...
r:Il
r:Il
....=
~
20
~
-'

......==
~
15

-'
~
~ 10
=
r.Il
5

0
0 2 4 6 8 10
Weeks from Calving
Figure 1. Incidence (diamonds) and prevalence (triangles) of subclinical ketosis in lactating
dairy cows. (Data from Duffield T: The placebo group of a large clinical trial involving 1010
Holstein dairy cows and 25 dairy farms. Unpublished, 1997.)

..
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 239

creased levels of butyric acid in the silage for certain years?7 Herd
differences found in these studies might also be partly attributed to
variations in the average herd parity. Several authors have reported a
higher prevalence of hyperketonemia with increasing lactation number.5,
21,22 Kauppinen found a similar parity relationship to hyperketonemia in
the highest ketone body class but not in the class chosen to define
subclinical ketosis. 45 Lack of an observed parity influence in this study
may have been a function of a liberal definition of subclinical ketosis,
because the prevalence (34%) was extremely high compared with other
studies. Steen et al found little difference in hyperketonemia between
primiparous and multiparous animals.72 This may have been caused by
a reduced detection rate of subclinical ketosis as a result of the milk
sampling frequency of every 2 months.
Body condition score (BCS) prior to calving is also an important
risk factor for subsequent development of subclinical ketosis during
lactation. Dohoo and Martin reported that a prolonged previous calving
interval increased the risk of subclinical ketosis in the subsequent lacta-
tion. 21 Duffield et al reported that fat (BCS of 4.0 or higher) cows had
both highest BHB concentrations postcalving and were at highest risk of
developing subclinical ketosis compared with cows in moderate and
thin body condition prior to calving. 24,25 Differences in the distribution
of fat, moderate, and thin cows between herds may also explain herd
differences in the occurrence of subclinical ketosis. Associations between
body condition pre calving and subclinical ketosis, clinical ketosis, and
multiple disease occurrence postcalving are illustrated in Figure 2.
Genetics and breed are other potential sources of herd variation.
Andersson and Emanuelson found that Swedish Red and White cows
had significantly higher milk acetone levels that Swedish Friesians. 5
Later studies observed the same breed differences with respect to the
incidence of clinical ketosis and the risk of hyperketonemia. 9, 40 Estimates
of the heritability of subclinical ketosis vary. Dohoo et al calculated the
heritability of clinical ketosis to be 0.32, but subclinical ketosis was
found to have no genetic component21; however, the authors suggested
that the interval between milk ketone body evaluations may have under-
estimated the heritability of the subclinical condition. Tveit et al reported
a heritability estimate for serum acetoacetate concentration of 0.11 with
a genetic correlation to milk yield of 0.87 in Norwegian cattle. 77 Manty-
saari et al estimated the heritability of clinical ketosis to be 0.07 to 0.09 in
Finnish Ayrshire cattle. 54 The heritability of clinical ketosis in Canadian
Holstein cows was recently estimated to be 0.09.80 The general conclusion
is that the heritability of both clinical and subclinical ketosis appears to
be low to moderate.
Season has been reported to influence the degree of hyperketonemia
in dairy cattle in several countries. A Norwegian study found increasing
levels of plasma acetoacetate with each month for calvings from August
to December.77 The authors suggested this observation may have been
caused by a systematic increase in body weight at freshening during the
calving season and an increasing tendency for fat cows at parturition. In
240 DUFFIELD

25~----------------------------------------~ 90

80
~
~ 20 70 ~
~ CD
CD Co)
Co) c
C CD
CD 60 "C
"C
"u 15 "u
.E
.E 50
..
tn
CD
en "en
ca 0
CD 40 CD
en 10 ~
is
Ci 30 Ci
u
u
:5 c:;
:5
(3 5 20 .c
::::s
en
10

o -+-----
Thin Good Fat
0
(104) (341 ) (62)
Body Condition Prior to Calving

Figure 2" Relationship between body condition score (BCS) precalving and the incidence
of subclinical ketosis (squares, serum beta-hydroxybutyrate ~ 1400 j.LmoI/L), clinical ketosis
(solid bars), and multiple illnesses (shaded bars, more than one periparturient disease) in
507 untreated cows. Thin = BCS precalving of ~ 3.0; good = BCS precalving between
3.25 and 3.75; fat = BCS precalving of ~ 4.0. (Data from Duffield T: 507 untreated Holstein
cows from 25 Canadian dairy farms. Unpublished, 1997.)

addition, the forage levels of butyric acid may have increased during
the same period. 77 Whitaker et al reported lower values of BHB in May
compared with other months of the year for dairy cows in the United
Kingdom. 81 Andersson and Emanuelson found levels of milk acetone to
be higher from October to February versus the remainder of the year in
Swedish dairy cows. S This is in agreement with Grohn et aI, who re-
ported that the risk of clinical ketosis was highest during the indoor
feeding season (September to May).37 Clinical ketosis was diagnosed
most often in cows that calved near the end of January; however, no
association between season and subclinical ketosis was found in a Cana-
dian study.21 The authors of the Canadian project suggested that there
may be certain factors that cause the expression of clinical ketosis that
are not associated with the subclinical condition21 ; however, it could also
be argued that the interval between samples in this study underesti-
mated the level of subclinical ketosis and reduced the opportunity to
observe a seasonal effect. In a subsequent Canadian study, both highest
BHB concentrations and highest subclinical ketosis prevalence occurred
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 241

in the summer months. 2s This may be the combined result of suppression

of intake caused by summer heat stress, changes in forages, and reduced
management intensity common to Ontario dairy farms from June to
September.

Subclinical Ketosis and Peri parturient Disease

Several investigations have evaluated the relationship between clini-

cal ketosis and other health variables, but few studies have attempted
to assess the associations between subclinical ketosis and periparturient
disease. Although subclinical and clinical ketosis are both part of the
same continuum, we can only assume the associations found for clinical
ketosis would be the same for the subclinical condition. Because there is
a paucity of information regarding the cause of subclinical ketosis and
peripartum conditions, the relationship between both clinical and sub-
clinical ketosis and other diseases is reviewed. Dohoo and Martin found
that cows with subclinical ketosis had an increased risk of metritis or
clinical ketosis 4 days later.2o The authors argued that because metritis
is a condition that normally develops at calving, subclinical ketosis is
more likely a result rather than a cause of metritis. Four subsequent
studies have found that the occurrence of retained placenta increased
the risk of clinical ketosis9, 15, 16, 38; however, the relationship between
metritis and ketosis is unclear. It seems likely that having a retained
placenta also increases the risk of a cow developing inflammation of the
uterus. 15, 16, 38, 61 The temporal association found in various studies be-
tween metritis and ketosis will depend on the timing of the diagnosis
of both conditions, because retained placenta increases the risk of both
diseases. The definition of metritis used by Dohoo and Martin included
cows with reproductive tract infections up to 65 days in milk, and
both conditions may have existed simultaneously20; however, subclinical
ketosis was identified prior to the diagnosis of metritis more often than
the reverse. A similar effect may have occurred in the study by Correa
et aI, where the definition of metritis used was any enlarged uterus
identified at rectal palpation and would have included metritis, endome-
tritis, and pyometra. 15 By contrast, both Rowlands et al and Grohn et al
found that early metritis diagnosed within 42 days postpartum (Grohn)
or endometritis identified within 25 days of calving (Rowlands) in-
creased the risk of ketosis. 38,64 Milk fever has also been linked with
clinical ketosis in several studies. 9, 16, 38
The relationship between displaced abomasum and ketosis has been
identified as bidirectional. l6, 38 That is, ketosis may be a cause of displace-
ment and abomasal displacement may lead to ketosis. Correa et al found
that ketosis increased the risk of abomasal displacement, but not the
reverse. 1S Dohoo and Martin could find no direct association between
the two conditions21; however, ketosis as an inciting or predisposing
cause of abomasal displacement can be further supported by some recent
Guelph research. Elevated BHB concentrations above 1000 J.Lmol/L in-
242 DUFFIELD

creased the likelihood of abomasal displacement. 34 Cows with concentra-

tions of BHB at or above 1400 J.1mol/L in the first 2 weeks post calving
were three times more likely to subsequently develop either clinical
ketosis or abomasal displacement. 23
Two studies have found a relationship between the diagnosis of
ketosis prior to identifying mastitis. 21,75 Mastitis increased the risk of
ketosis in Finnish Ayrshires. 38 Hyperketonemic cows with BHB blood
levels above 1400 J.1mol/L were found to suffer a more severe experi-
mental mastitis than normal COWS. 48 A possible explanation for ketosis
increasing the risk of mastitis might be a potential inhibitory effect of
ketone bodies on the function of white blood cells. 83 There may be some
important immune function implications associated with decreased en-
ergy balance and subclinical ketosis. A decreased risk of cows devel-
oping more than one disease postcalving when treated prophylactically
with a monensin-controlled release capsule has been reported. 27 This
finding may be the result of improved immune function through im-
proved energy metabolism; however, this area requires more research.

Subclinical Ketosis and Milk Production

In general, there is consensus that a negative association between

hyperketonemia and milk production exists; however, there are conflict-
ing reports. In one study, the loss of production associated with a
positive milk ketone test was 1.0 to 1.4 kg of milk per day.20 This
represented 4.40/0 to 6.00/0 of the mean test day milk production. In this
study, both the milk samples and the milk weights were obtained on the
same day, and ketolactia and milk production were evaluated concur-
rently. Test-day milk production was negatively correlated with milk
acetone levels in four separate Scandinavian projects.5, 40, 57,72 By contrast,
Kauppinen showed a significant positive correlation between the BHB
and acetoacetate concentrations in blood and milk yield. 44 Although
somewhat unclear, this study appears to be a cross-sectional type of
design. Blood was sampled once from each of 662 cows that ranged
from nonlactating to greater than 101 days in milk. Because milk re-
cording was conducted monthly, blood samples were not likely to be
obtained on the same day as milk yield determination. Kauppinen
subsequently reported that subclinically ketotic cows had significantly
higher annual milk yields than nonketotic COWS. 46
Fat-corrected milk yield increased with increasing levels of plasma
acetoacetate in two groups of half-sister related heifers in a Norwegian
feeding trial, but the average plasma acetoacetate levels did not exceed
100 J.1mol/L (760 J.1mol/L BHB) until after 3 weeks postpartum, when
peak milk yield was attained. 42 Milk fat percent has been shown to be
significantly higher in cows with subclinical ketosis,57 and the production
of fat-corrected milk may be higher in subclinically ketotic cows. Both
test-day milk yield and fat-corrected milk were lower in Swedish cows
that were hyperketonemic. 40
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 243

Herdt et al found higher concentrations of BHB in higher-producing

cows, but individual milk tests preceded blood measurement for BHB.43
It is possible that higher milk yields put cows at increased risk of
developing subclinical ketosis. Increased milk production may be associ-
ated with increased fat mobilization and a greater risk of hyperketo-
nemia. 52 Studies measuring both ketone body concentrations and milk
production on the same day should identify any potential negative
impacts of hyperketonemia on milk output. Trials in which milk mea-
surements precede ketone evaluation could identify high-producing
cows that later become subclinically ketotic. Because of the difference in
timing of ketone body measurement and milk yield determination, a
positive correlation between milk production and hyperketonemia
would be observed.
In a review of several observational studies, Erb concluded that
higher milk yields in the previous lactation did not increase the risk of
ketosis in the subsequent lactation. 31 One Finnish study determined
increased previous milk yield to be a risk factor for ketosis,38 but this
association was not identified in two North American projects. 16, 21 Re-
gardless of the threshold chosen for hyperketonemia, the incidence of
subclinical ketosis should be considerably higher than that of clinical
ketosis. Assuming that higher-producing cows are more likely to be
hyperketonemic, a larger proportion of subclinically ketotic cows could
mute any association between clinical ketosis and previous lactation
milk yield. One study that evaluated test-day milk production and
clinical ketosis in over 60,000 Finnish Ayrshires demonstrated a lactation
curve depression associated with ketosis and a loss of 44.3 kg of milk18;
however, cows with ketosis had 141.1 kg more 305-day milk yield than
normal COWS. 18 Inverted milk curves in hyperketonemic cows were also
noted in a large Swedish study.40 Most of the loss in milk production
occurred in the first 100 days of lactation and amounted to 328 kg loss
in fat-corrected milk yield over 200 days.

Subclinical Ketosis and Milk Components

Milk fat and milk protein are significantly altered in hyperketo-

nemia. Milk fat percentage was increased in subclinically ketotic COWS56,57
and in cows with clinical ketosis.71 Mean annual milk fat yield was
significantly higher in both subclinically ketotic and clinically ketotic
cows compared with normal COWS. 46 The association between milk fat
and hyperketonemia is, presumably, because of increased availability of
BHB and fatty acids for milk fat synthesis. It is unclear whether increased
concentrations of circulating ketone bodies cause increased milk fat, or
whether cows that are prone to higher milk fat yields are more suscepti-
ble to subclinical ketosis. Milk protein percent has been reported to be
lower in cows with subclinical ketosis. 56,57 This may be the result of a
reduced energy supply, because milk protein percent is positively associ-
ated with net energy balance. 36
244 DUFFIELD

Subclinical Ketosis and Reproductive Performance

Increasing the degree of negative energy balance in early lactation

has been shown to increase the interval from calving to first ovulation. 13
Butler and Smith suggested that cows with a longer interval from calving
to first ovulation experience a decrease in pregnancy rate at first ser-
vice,13 because conception rate is related to the number of ovulatory
cycles that occur prior to insemination. 73, 84 Because hyperketonemia is a
sign of disturbed energy metabolism, many authors have investigated
the relationship between subclinical ketosis and reproductive perfor-
mance. No effect of either subclinical or clinical ketosis on individual
cow fertility was found in two studies5, 46; however, significant correla-
tions between the herd prevalence of hyperketonemia and herd mean
intervals from calving to first service and calving to last service have
been noted. 5 A link between subclinical ketosis and the increased inci-
dence of cystic ovaries has also been reported. 5,20 A significant inverse
relationship between milk fat percentage and first insemination preg-
nancy rates has been identified. 49 Miettenen and Setala found an in-
creased interval from calving to conception in cows with high milk yield
and high fat yield. 57 The associations between fertility and increased fat
and milk yield do not necessarily imply a relationship between impaired
fertility and hyperketonemia. The duration of either clinical or subclini-
cal ketosis may be too short to exert a negative effect on calving inter-
va1. 46 Whitaker et al found that cows with a better energy status at 14
days postpartum had a reduced interval from calving to the onset of
cyclicity and fewer services per conception. 82 No effect was observed
when energy status was evaluated at 21 days postpartum or at first
service. This study was only conducted on 24 cows within one herd. It
is not clear if hyperketonemia truly has a negative effect on reproduction
or whether hyperketonemia and impaired reproductive performance are
simply consequences of a prolonged negative energy balance associated
with increased milk production.

PREVENTION OF SUBCLINICAL KETOSIS

Because ketosis occurs in early lactation, recommendations for pre-

vention have focused on the nutritional management of the dry and
transition cow. Detailed recommendations for nutrition during the dry
period can be found elsewhere. 59 It is a common recommendation to
divide the dry period into two feeding groups: far-off and close-up.62
Typically, far-off diets follow National Research Council (NRC) guide-
lines for dry cows, but there are no NRC guidelines for close-up (pre-
fresh) cows. The close-up diet is usually balanced according to recom-
mendations that are halfway between the dry cow and early-lactation
cow and should be fed starting at least 3 weeks before expected calving.59
The goals of the transition diet are specifically designed to prevent
subclinical ketosis by maximizing dry matter intake and providing ade-

II
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 245

quate energy density. 59 Avoidance of ketogenic feedstuffs 77 and increased

frequency of feeding concentrates4,40 have been advocated as preventive
measures against subclinical ketosis. Avoiding overconditioning of cows
in late lactation and the early dry period, as well as lead feeding with
concentrates about 3 weeks prior to calving, have also been suggested
as aids in prophylaxis. 4,51 See the articles by Gerloff and Oetzel elsewhere
in this issue for a more complete discussion of feeding transition cows.
In addition to good nutrition, certain feed additives have been
found beneficial in reducing subclinical ketosis when administered pro-
phylactically. Niacin fed prior to calving at the rate of 6 to 12 grams per
day may be helpful in reducing blood levels of BHB.28, 33 Propylene glycol
has been used successfully for the prevention of subclinical ketosis. 29, 65 A
dose of propylene glycol of 1 L per day as an oral drench for 9 days
prior to calving decreased BHB and nonesterified fatty acids and in-
creased glucose concentrations.74 It appears that a bolus of propylene
glycol is necessary for maximum effect, because mixing in a total mixed
ration is not as efficacious as either an oral drench or when mixed with
a small quantity of grain. 14 Schultz reported that sodium propionate
could be given to prevent clinical ketosis in dairy cattle. 67 Propylene
glycol requires repeated daily oral administration, and sodium propio-
nate may reduce feed intake. 66 Ionophores have been proposed as poten-
tial prophylactic agents for reducing hyperketonemia. 52, 78 In contrast
to propylene glycol and sodium propionate, ionophores are relatively
inexpensive and much easier to administer. Ionophores are not approved
for use in lactation-age dairy cows in the United States.
The gluconeogenic potential of monensin has attracted researchers
to investigate its possible role as an antiketogenic agent in dairy cattle.
Rogers and Hope-Cawdery first described the beneficial effects of mo-
nensin for reducing the incidence of ketosis in a herd with a clinical
ketosis problem. 63 The antiketogenic properties of monensin were later
investigated in a Canadian trial involving two levels of monensin and
three groups of 12 Holstein COWS. 66 Monensin included at 30 grams per
ton of total ration (high group) decreased the incidence of subclinical
ketosis and significantly reduced blood BHB levels in the first 3 weeks
postpartum. 66 The incidence of subclinical ketosis, defined as total blood
ketone bodies over 9 mg/ dL (900 fJ,.mol/L), was reduced by 50%, and
blood BHB levels were reduced by 40% for the high-monensin group.66
Based on the average feed intakes observed in this trial, the low-monen-
sin group received approximately 208 mg of monensin per day, and the
high group 399 mg per day.66 Monensin treatment commencing at 2 to 4
weeks prior to calving reduced serum BHB and nonesterified fatty acids
(NEFAs) in lactating dairy cows during the first 28 days postpartum
when monensin was fed at 300 or 450 mg / d, but not by a daily dose of
150 mg / d of monensin.76 Serum glucose was not influenced by monensin
feeding. Australian cows treated with a monensin controlled-release
capsule during the first week postcalving had significantly lower plasma
BHB levels and tended to have higher glucose concentrations than
controls.1 A controlled-release capsule that delivers 335 mg of monensin
246 DUFFIELD

sodium per day for 95 days reduced the incidence of subclinical ketosis
by 50% and also decreased the duration of the condition when it was
administered 3 weeks prior to expected calving. 24 Monensin and other
ionophores are not approved for use in dairy cows of lactation age in
the United States.

ECONOMIC COSTS OF SUBCLINICAL KETOSIS

Costs of subclinical ketosis include both lost potential milk produc-

tion and increased occurrence of periparturient disease. These costs will
vary from farm to farm depending on the incidence of disease, distribu-
tion of risk factors, and nutritional management of the dry cow. The
most closely linked diseases occurring subsequent to subclinical ketosis
are displaced abomasum and clinical ketosis. Estimates of milk produc-
tion loss range from 300 to 450 kg for a lactation. 20,40 These losses must
be weighed against the cost of any prophylactic measure. The results
from a recent study involving the use of ionophores for preventing
subclinical ketosis are used to illustrate the potential costs associated
with the disease. Administration of a monensin controlled-release cap-
sule pre calving reduced the incidence of subclinical ketosis by 500/0,25
clinical ketosis by 500/0, abomasal displacement by 40%, and multiple
illness by 400/0 (more than one disease).27 The milk production response
depended on body condition and was 0.85 kg / d at peak lactation in
cows with a precalving BCS of 3.25 to 3.75, and was 1.2 kg/ d for the
first 90 days of lactation in fat cows (BCS of 4.0 or greater).26 No milk
production response was noted in thin cows, presumably because they
had the lowest BHB concentrations and were at decreased risk of subclin-
ical ketosis. A spreadsheet has been developed to estimate the impact of
these findings on dairy herds. Table 2 outlines two herd examples taken
from the spreadsheet. Herd A has an average distribution of body
condition scores and an average incidence of clinical disease for the 25
herds that participated in the study. In this example, the benefit of using
the prophylactic treatment is 3 to 1 and would be cost-effective in this
herd. Herd B has more thin cows but less metabolic disease problems.
There is still a positive payback for prophylaxis, but it is less than 2 to
1, and in this example the herd owner may want to choose another
method of prevention or use selective prevention based on body condi-
tion score. If the preventive benefit of ionophore use is gained solely
from the prevention of subclinical ketosis, then the cost of subclinical
ketosis to these two dairies can be assumed to be CAN $4800 and CAN
$2400 for herd A and herd B, respectively. Because monensin has been
shown to reduce the incidence of subclinical ketosis by 50%, the cost of
subclinical ketosis using these two examples has an approximate range
of CAN $50 to CAN $100 per cow per lactation. Benefits and costs will
vary according to several variables, including cost of disease, cost of
treatment, price of milk, disease incidence, and distribution of body
condition in the herd.
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 247

Table 2. ESTIMATED ECONOMICS OF PREVENTING SUBCLINICAL KETOSIS USING

A MONENSIN CONTROLLED-RELEASE CAPSULE IN TWO EXAMPLE DAIRY HERDS*

Variable Herd A Herd B

Herd Data
Calvings per year 100 100
Number of cows with BCS 2: 4.0 prior to calving (fat) 15 5
Number of cows with BCS s 3.0 prior to calving (thin) 5 20
LDA cases per year 5 2
Clinical ketosis cases per year 2 2
Multiple illness cases (cows with more than one 13 4
periparturient disease) per year
Prices and Costs ($)
Cost of monensin capsule 16 16
Price of milk/hectoliter 54 54
Cost of LDA 500 500
Cost of ketosis 200 200
Cost of multiple illness 500 500
Health Savings ($)
LDA 1000 400
Ketosis 200 200
Multiple illness 1625 500
Total 2825 1100
Milk Production Benefit
Thin cows 0 0
BCS 3.25 in 3.75 ($) 1100 1033
Fat cows ($) 875 292
Total milk benefit ($) 1975 1325
Cost of treatment ($) 1600 1600
Overall benefit ($) 4800 2425
Benefit:cost ratio 3.0 1.5

*Assumptions: (1) monensin controlled-release capsule reduces ketosis by 50%, left displaced
abomasum (LOA) by 40%, and multiple illness by 40%; (2) LOA and ketosis account for 25% of
multiple illness (therefore adjusted down); (3) no production response in thin cows; (4) + 0.85 kg/ d for
30 days in cows with BCS 3.25 to 3.75 prior to calving; (5) + 1.2 kg/ d for 90 days in cows with
BCS :2':4.0 prior to calving; (6) cost of disease based on estimates by Dr. Chuck Guard (personal
communication, 1999).
These costs include treatment, culling and replacement, and milk production loses. Cost of multiple
illness assumed to be at least the same as LOA.

MONITORING PROGRAMS

Dairy herds could benefit from a routine monitoring program for

subclinical ketosis. Reasons for monitoring subclinical ketosis may in-
clude evaluation of prophylactic measures, monitoring the transition
ration, and identification of animals for treatment prior to development
of clinical disease. Before a program is instituted, the veterinarian and
farm manager need to know what the average incidence is for the herd
so that a reasonable and achievable target can be set. In the recent trial
conducted at the University of Guelph, the mean incidence of subclinical
ketosis (BHB of 1400 f-Lmol/L or greater) in untreated cows was 54%.
The range across 25 herds was 8% to 80%. The four highest herds had
incidence rates above 65% and also had the largest milk production
response to prophylactic treatment. An appropriate monitoring program
248 DUFFIELD

would assess cows in the first 2 weeks of lactation, because this is the
time of peak incidence. Also, identification of positive cows earlier in
lactation might allow time for prevention of subsequent clinical disease.
Choice of an appropriate test is difficult. Ketolac has been shown to be
a very useful test; however, it is presently unavailable in North America
and is considerably more expensive than the currently available nitro-
prusside tests for ketone bodies. The nitroprusside-based milk ketone
tests have excellent specificity, so there is little danger of false positives;
however, these tests lack sensitivity, limiting their usefulness as screen-
ing tests. The nitroprusside test in urine provides almost a zero probabil-
ity of a false negative (ideal for screening), but there is close to a 500/0
chance of a false positive. The prevalence of disease will also alter the
predictive values of the various tests. Predictive values for three cowside
tests used within hypothetical herds having 200/0, 40 % , and 60% preva-
lence of subclinical ketosis in the first 2 weeks of lactation are presented
in Table 3. The Ketolac test performs well across all prevalence levels.
The Ketocheck test tends to give better predictive values at lower preva-
lences, and the urine ketone test yields higher predictive values at higher
prevalences. The predictive value should be compared with the prior
prevalence of disease (true prevalence), and the difference between the
two numbers is the information gained from the test. For example, at
20% prevalence, there is a 200/0 chance that any animal selected prior to
testing would have subclinical ketosis. Using the urine ketone test, a
positive result would gain only 180/0 (predictive value of a positive test
result: 38 % ), whereas the Ketocheck test would gain 70%.
It is most likely that in screening a group of fresh cows, there would
be two possible actions resulting from the test. One action might be to
treat positive animals with the goal to prevent subsequent development
of clinical disease. In this case, a high predictive value of a positive test
is desired so that normal animals are not unnecessarily treated. The
second action might be to compare the percent of positive reactors to a
goal for determining the effectiveness of either the transition ration or
some prophylactic measure in reducing the incidence of subclinical
ketosis. In this situation, the apparent prevalence is the parameter that
actually would be used. Note from Table 3 that the urine ketone test
would substantially overestimate the prevalence of subclinical ketosis,
whereas the Ketocheck test would grossly underestimate the prevalence.
This does not preclude these tests from being used; however, the impact
of the inherent sensitivity and specificity of the test must be remembered
when establishing goals and intervention thresholds.

CONCLUSIONS

Subclinical ketosis is an important metabolic disease in early-lacta-

tion dairy cattle that is associated with milk production losses and
several periparturient diseases. Because the cause is related to the ho-
meorhetic drive for high levels of milk production, the incidence of this
 Table 3. USE OF COWSIDE KETONE TESTS IN SCREENING PROGRAMS FOR IDENTIFYING SUBCLINICAL KETOSIS
60%
20% Prevalence 40% Prevalence Prevalence
Apparent Apparent Apparent
PV+ PV- Prevalence PV+ PV- Prevalence PV+ PV- Prevalence
Test (%) (%) (%) (%) (%) (%) (%) (%) (%)
Ketolac using 100 ~mol/L 62 93 23 81 83 35 91 68 48
Ketocheck at 1400 ~mol/L 90 86 8 96 70 16 98 51 23
BHB
Utrecht urine 38 100 53 62 100 65 78 100 76
Ketone 1400 ~mol/L BHB

PV + = predictive value of a positive test result; PV - = predictive value of a negative test result.

~
\C
250 DUFFIELD

condition is likely to remain important as the dairy industry strives for

higher levels of milk yield. Prevention of this disease is primarily
achieved through proper nutritional programs for the dry and early-
lactation cow. In addition, management of body condition score to re-
duce overconditioned or fat cows prior to subsequent calving is an
important management factor for reducing subclinical ketosis incidence.
Certain feed additives, such as niacin, propylene glycol, and especially
ionophores, are also helpful in managing this condition, although iono-
phores cannot be used as feed additives for dry or lactating dairy cows
in the United States. Testing programs for individual identification and
herd monitoring should focus on the first few weeks of lactation. Veteri-
narians must bear in mind the test performance characteristics when
devising herd monitoring programs for subclinical ketosis. Implementa-
tion of a herd monitoring program for subclinical ketosis would be a
useful addition to many herd health programs.

References

1. Abe N, Lean IJ, Rabiee A, et al: Effects of sodium monensin on reproductive perfor-
mance of dairy cattle: II. Effects on metabolites in plasma, resumption of ovarian
cyclicity and oestrus in lactating cows. Aust Vet J 71:277-282, 1984
2. Adler JH, Roberts SJ, Steel RGD: The relation between reactions to the Ross test on
milk and urine and the degree of ketonemia in dairy cows. Cornell Veterinarian
47:101-111, 1956
3. Andersson L: Concentrations of blood and milk ketone bodies, blood isopropanol and
plasma glucose in dairy cows in relation to the degree of hyperketonemia and clinical
signs. Zentralbl Veterinarmed A 31:683-693, 1984
4. Andersson L: Subclinical ketosis in dairy cows. Vet Clin North Am Food Anim Pract
4:233-248, 1988
5. Andersson L, Emanuelson U: An epidemiological study of hyperketonaemia in Swed-
ish dairy cows; Determinants and the relation to fertility. Preventive Veterinary Medi-
cine 3:449-462, 1985
6. Baird GD, Hibbitt KG, Hunter GD, et al: Biochemical aspects of bovine ketosis.
Biochem J 107:683-689, 1968
7. Baird DG: Ruminant ketosis. Biochem Soc Trans 9:348-349, 1981
8. Baird DG: Primary ketosis in the high-producing dairy cow: Clinical and subclinical
disorders, treatment, prevention and outlook. J Dairy Sci 65:1-10, 1982
9. Bendixen PH, Vilson B, Ekesbo BI, et al: Disease frequencies in dairy cows in Sweden:
IV. Ketosis. Preventive Veterinary Medicine 5:99-109, 1987
10. Bergman EN: Hyperketonemia-ketogenesis and ketone body metabolism. J Dairy Sci
54:936-948, 1971
11. Blood DC, Radostits OM: Veterinary Medicine. Toronto, Bailliere & Tindall, 1989,
pp 1128-1138
12. Bruss ML: Ketogenesis and ketosis. In Kaneko JJ (ed): Clinical Biochemistry of Domes-
tic Animals, ed 4. Toronto, Academic Press, 1989, pp 86-105
13. Butler WR, Smith RD: Interrelationships between energy balance and postpartum
reproductive function in dairy cattle. J Dairy Sci 72:767-783, 1989
14. Christensen JO, Rasmussen FE, Grummer RR: Influence of propylene glycol delivery
method on plasma metabolites of feed restricted cattle. J Dairy Sci 78 (suppl1):240, 1995
15. Correa MT, Erb H, Scarlett J: Path analysis for seven postpartum disorders of Holstein
cows. J Dairy Sci 76:1305-1312, 1993
16. Curtis CR, Erb HN, Sniffen CJ, et al: Path analysis of dry period nutrition, postpartum
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 251

metabolic and reproductive disorders, and mastitis in Holstein cows. J Dairy Sci
68:2347-2360, 1985
17. Custer EM, Myers JL, Poffenbarger PL, et al: The storage stability of 3-hydroxybutyrate
in serum, plasma, and whole blood. Am J Clin Pathol 80:375-380, 1983
18. Detilleux JC, Grohn YT, Quass RL: Effects of clinical ketosis on test day milk yields in
Finnish Ayrshire cattle. J Dairy Sci 77:3316-3323, 1994
19. Dirksen G, Breitner W: A new quick-test for semiquantitative determination of beta-
hydroxybutyric acid in bovine milk. Journal of Veterinary Medicine A 40:779-784, 1993
20. Dohoo IR, Martin SW: Subclinical ketosis: Prevalence and associations with production
and disease. Can J Comp Med 48:1-5, 1984
21. Dohoo IR, Martin SW: Disease, production and culling in Holstein-Friesian cows III:
Disease and production as determinants of disease. Preventive Veterinary Medicine
2:671-690, 1984
22. Duffield TF, Kelton DF, Leslie KE, et al: Use of test day milk fat and milk protein to
predict subclinical ketosis in Ontario dairy cattle. Can Vet J 38:713-718, 1997
23. Duffield TF: Effects of a monensin controlled release capsule on energy metabolism,
health, and production in lactating dairy cattle. Thesis dissertation, Guelph, Ontario,
University of Guelph, 1997
24. Duffield TF, Sandals D, Leslie KE, et al: Effect of prepartum administration of a
monensin controlled release capsule on postpartum energy indicators in lactating dairy
cattle. J Dairy Sci 81:2354-2361, 1998
25. Duffield TF, Sandals D, Leslie KE, et al: Efficacy of monensin for the prevention of
subclinical ketosis in lactating dairy cows. J Dairy Sci 81:2866-2873, 1998
26. Duffield TF, Leslie KE, Sandals D, et al: Effect of prepartum administration of a
monensin controlled release capsule on milk production and milk components in early
lactation. J Dairy Sci 82:272-279, 1999
27. Duffield TF, Leslie KE, Sandals D, et al: Effect of prepartum administration of a
monensin controlled release capsule on cow health and reproductive performance. J
Dairy Sci 82:2377-2384, 1999
28. Dufva GS, Bartley EE, Dayton AD, et al: Effect of niacin supplementation on milk
production and ketosis of dairy cattle. J Dairy Sci 66:2329-2336, 1983
29. Emery RS, Burg N, Brown LD, et al: Detection, occurrence, and prophylactic treatment
of borderline ketosis with propylene glycol feeding. J Dairy Sci 47:1074-1079, 1964
30. Emery RS, Bell JW, Thomas JW: Benefits derived from routine testing for milk ketones.
J Dairy Sci 51:867-868, 1969
31. Erb HN: Interrelationships among production and clinical disease in dairy cattle: A
review. Can Vet J 28:326-329, 1987
32. Faulkner A, Peaker M: Reviews of the progress of dairy science: Secretion of citrate
into milk. J Dairy Res 49:159-169, 1982
33. Fronk TJ, Schultz LH: Oral nicotinic acid as a treatment for ketosis. J Dairy Sci
62:1804-1807, 1979
34. Geishauser T, Leslie K, Duffield T, et al: Evaluation of aspartate aminotransferase
activity and ~-hydroxybutyrate concentration in blood as tests for left displaced abo-
masum in dairy cows. Am J Vet Res 58:1216-1220, 1997
35. Geishauser T, Leslie K, Kelton D, et al: Evaluation of five cowside tests for use with
milk to detect subclinical ketosis in dairy cows. J Dairy Sci 81:438---443, 1998
36. Grieve DG, Korver S, Rijpkema YS, et al: Relationship between milk composition and
some nutritional parameters in early lactation. Lact Prod Sci 14:239-254, 1986
37. Grohn Y, Thompson JR, Bruss ML: Epidemiology and genetic basis of ketosis in
Finnish Ayrshire cattle. Preventive Veterinary Medicine 3:65-77, 1984
38. Grohn YT, Erb HN, McCulloch CE, et al: Epidemiology of metabolic disorders in dairy
cattle: Association among host characteristics, disease, and production. J Dairy Sci
72:1876-1885, 1989
39. Gustafsson AH: Acetone and urea concentration in milk as indicators of the nutritional
status and the composition of the diet of dairy cows. Thesis Dissertation, Uppsala,
Swedish University of Agricultural Science, 1993, pp 1-48
40. Gustafsson AH, Andersson L, Emanuelson U: Effect of hyperketonemia, feeding fre-
252 DUFFIELD

quency and intake of concentrate and energy on milk yield in dairy cows. Animal
Production 56:51-60, 1993
41. Gustafsson AH, Andersson I, Emanuelson U: Influence of feeding management, con-
centrate intake and energy intake on the risk of hyperketonaemia in Swedish dairy
herds. Preventive Veterinary Medicine 22:237-248, 1995
42. Halse K, Tveit B: Prefeeding plasma acetoacetate and glucose in health, lactating
heifers: Variations related to milk yield, metabolic balances and stage of lactation. Acta
Vet Scand 35:243-255, 1994
43. Herdt TH, Stevens JB, Olson WG, et al: Blood concentrations of f3-hydroxybutyrate in
clinically normal Holstein-Friesian herds and in those with a high prevalence of clinical
ketosis. Am J Vet Res 12:503-506, 1981
44. Kauppinen K: Correlation of whole blood concentrations of acetoacetate, B-hydroxybu-
tyrate, glucose and milk yield. Acta Vet Scand 24:337-348, 1983
45. Kauppinen K: Prevalence of bovine ketosis in relation to number and stage of lactation.
Acta Vet Scand 24:349-361, 1983
46. Kauppinen K: Annual milk yield and reproductive performance of ketotic and non-
ketotic dairy cows. Zentralbl Veterinarmed A 31:694-704, 1984
47. Kelly JM: Changes in serum f3-hydroxybutyrate concentrations in dairy cows kept
under commercial farm conditions. Vet Rec 101:499-502, 1977
48. Kremer WDJ, Noordhuizen-Stassen EN, Grommers FJ, et al: Severity of experimental
Escherichia coli mastitis in ketonemic and nonketonemic dairy cows. J Dairy Sci 76:3428-
3436, 1993
49. Kristula MA, Reeves M, Redlus H, et al: A preliminary investigation of the association
between the first postpartum milk fat test and first insemination pregnancy rates.
Preventive Veterinary Medicine 23:94-100, 1995
50. Kronfeld DS, Raggi F, Ramberg CF: Mammary blood flow and ketone body metabolism
in normal, fasted and ketotic cows. Am J Physiol 215:218-227, 1968
51. Lean IJ, Bruss ML, Baldwin RL, et al: Bovine ketosis: A review: I. Epidemiology and
pathogenesis. Veterinary Bulletin 61:1209-1218, 1991
52. Lean IJ, Bruss ML, Baldwin RL, et al: Bovine ketosis: A review. II. Biochemistry and
prevention. Veterinary Bulletin 62:1-14, 1992
53. Manston R, Rowlands GJ, Little W, et al: Variability of the blood composition of dairy
cows in relation to time of day. Journal of Agricultural Science 96:593-598, 1981
54. Mantysaari EA, Grohn YT, Quaas RL: Clinical ketosis: Phenotypic and genetic correla-
tions between occurrences and with milk yield. J Dairy Sci 74:3985-3993, 1991
55. Marstorp P, Anfalt T, Anderson L: Determination of oxidized ketone bodies in milk
by flow injection analysis. Analytic Chimica Acta 149:281-289, 1983
56. Miettenen PVA: Relationship between milk acetone and milk yield in individual cows.
Journal of Veterinary Medicine A 41:102-109, 1994
57. Miettinen PVA, Setala JJ: Relationships between subclinical ketosis, milk production
and fertility in Finnish dairy cattle. Preventive Veterinary Medicine 17:1-8, 1993
58. Nielen M, Aarts MGA, Jonkers GM, et al: Evaluation of two cowside tests for the
detection of subclinical ketosis in dairy cows. Can Vet J 35:229-232, 1994
59. Oetzel GR: Dairy: Nutrition management. Nutritional management of dry dairy cows.
Compend Contin Educ Pract Vet Food Animal 20:391-396, 1998
60. Palmquist DL, Davis CL, Brown RE, et al: Availability and metabolism of various
substrates in ruminants: V. Entry rate into the body and incorporation into milk fat of
D( - )B-hydroxybutyrate. J Dairy Sci 52:633-638, 1969
61. Peeler EJ, Otte MJ, Esslemont RJ: Inter-relationships of periparturient diseases in dairy
cows. Vet Rec 134:129-132, 1994
62. Radostits OM, Leslie KE, Fetrow J: Dairy cattle nutrition. In Herd Health: Food Animal
Production Medicine, ed 2. Philadelphia, WB Saunders, 1994, pp 277-300
63. Rogers PAM, Hope-Cawdery MJ: Monensin, ketosis and nitrate toxicity in cows. Vet
Rec 106:311-312, 1980
64. Rowlands GJ, Lucey S, Russell AM: Susceptibility to disease in the dairy cow and its
relationship with occurrences of other diseases in the current or preceding lactation.
Preventive Veterinary Medicine 4:223-234, 1986
65. Sauer FD, Erfle JD, Fisher LJ: Propylene glycol and glycerol as a feed additive for
 SUBCLINICAL KETOSIS IN LACTATING DAIRY CATTLE 253

lactating dairy cows: An evaluation of blood metabolite parameters. Canadian Journal

of Animal Science 53:265-271, 1973
66. Sauer FD, Kramer JKG, Cantwell WJ: Antiketogenic effects of monensin in early
lactation. J Dairy Sci 72:436-442, 1989
67. Schultz LH: Use of sodium propionate in the prevention of ketosis in dairy cattle. J
Dairy Sci 41:160-168, 1958
68. Schultz LH, Myers M: Milk test for ketosis in dairy cows. J Dairy Sci 42:705-710, 1959
69. Schultz LH: Management and nutritional aspects of ketosis. J Dairy Sci 54:962-973, 1971
70. Simensen E, Halse K, Gillund P, et al: Ketosis treatment and milk yield in dairy cows
related to milk acetoacetate levels. Acta Vet Scand 31:433-440, 1990
71. Solbu H: Disease recording in Norwegian dairy cattle: I. Disease incidences and
non-genetic effects on mastitis, ketosis and milk fever. Z Tierz Zuchtungsbiology
100:139-157, 1983
72. Steen A, Osteras 0, Gronstol H: Evaluation of additional acetone and urea analyses,
and of the fat-Iactose-quotient in cow milk samples in the herd recording system in
Norway. Journal of Veterinary Medicine A 43:181-191, 1996
73. Stevenson JS, Call EP: Influence of early estrus, ovulation, and insemination on fertility
in postpartum Holstein cows. Theriogenology 19:367-375, 1983
74. Struder VA, Grummer RR, Bertics SJ: Effect of prepartum propylene glycol administra-
tion on periparturient fatty liver in dairy cows. J Dairy Sci 76:2931, 1993
75. Syvajarvi J, Saloniemi H, Grohn Y: An epidemiological and genetic study on registered
diseases in Finnish Ayrshire cattle. Acta Vet Scand 27:223-234, 1986
76. Thomas EE, Poe SE, McGuffey RK, et al: Effect of feeding monensin to diary cows on
milk production and serum metabolites during early lactation. J Dairy Sci 76(suppl
1):280, 1993
77. Tveit B, Lingaas F, Svendsen M, et al: Etiology of acetonemia in Norwegian cattle:
1. Effect of ketogenic silage, season, energy level, and genetic factors. J Dairy Sci
75:2421-2432, 1992
78. Tyler JW, Wolfe DF, Maddox R: Clinical indications for dietary ionophores in rumi-
nants. Compendium of Continuing Education 14:989-993, 1992
79. Tyopponen J, Kauppinen K: The stability and automatic determination of ketone
bodies in blood samples taken in field conditions. Acta Vet Scand 21:55-61, 1980
80. Uribe HA, Kennedy BW, Martin SW, et al: Genetic parameters for common health
disorders of Holstein cows. J Dairy Sci 78:421-430, 1995
81. Whitaker DA, Kelly JM, Smith EJ: Subclinical ketosis and serum beta-hydroxybutyrate
levels in dairy cattle. Br Vet J 139:462-463, 1983
82. Whitaker DA, Smith EJ, da Rosa GO, et al: Some effects of nutrition and management
on the fertility of dairy cattle. Vet Rec 133:61-64, 1993
83. White F, Rattray EAS: The in vitro effect of ketone bodies on the growth curves of
mastitis organisms in milk. J Comp Pathol 78:179-188, 1968
84. Whitmore HL, Tyler WJ, Casida LE: Effects of early postpartum breeding in dairy
cattle. J Anim Sci 38:339-346, 1974

Address reprint requests to

Todd Duffield, DVM, DVSc
Department of Population Medicine
Ontario Veterinary College
University of Guelph
Guelph, Ontario
Canada N1G 2W6

e-mail: [email protected]