Human Nutrition and Metabolism

␤-Hydroxy-␤-Methylbutyrate (HMB) Supplementation in Humans Is Safe and

May Decrease Cardiovascular Risk Factors
S. Nissen,*1 R. L. Sharp,* L. Panton,* M. Vukovich,† S. Trappe‡ and J. C. Fuller, Jr.**
*Iowa State University, Ames, IA, †Wichita State University, Wichita, KS, ‡Ball State University, Muncie, IN
and **Metabolic Technologies Incorporated, Iowa State University Research Park, Ames, IA

ABSTRACT The leucine metabolite, ␤-hydroxy-␤-methylbutyrate (HMB) enhances the effects of exercise on
muscle size and strength. Although several reports in animals and humans indicate that HMB is safe, quantitative
safety data in humans have not been reported definitively. The objective of this work was to summarize safety data
collected in nine studies in which humans were fed 3 g HMB/d. The studies were from 3 to 8 wk in duration,
included both males and females, young and old, exercising or nonexercising. Organ and tissue function was
assessed by blood chemistry and hematology; subtle effects on emotional perception were measured with an
emotional profile test (Circumplex), and tolerance of HMB was assessed with a battery of 32 health-related
questions. HMB did not adversely affect any surrogate marker of tissue health and function. The Circumplex
emotion profile indicated that HMB significantly decreased (improved) one indicator of negative mood (Unactivated
Unpleasant Affect category, P ⬍ 0.05). No untoward effects of HMB were indicated. Compared with the placebo,
HMB supplementation resulted in a net decrease in total cholesterol (5.8%, P ⬍ 0.03), a decrease in LDL
cholesterol (7.3%, P ⬍ 0.01) and a decrease in systolic blood pressure (4.4 mm Hg, P ⬍ 0.05). These effects of
HMB on surrogate markers of cardiovascular health could result in a decrease in the risk of heart attack and stroke.
In conclusion, the objective data collected across nine experiments indicate that HMB can be taken safely as an
ergogenic aid for exercise and that objective measures of health and perception of well-being are generally
enhanced. J. Nutr. 130: 1937–1945, 2000.

KEY WORDS: ● leucine ● cholesterol ● blood pressure ● toxicity ● humans

␤-Hydroxy-␤-methylbutyrate (HMB)2 is a common dietary 1983) and, at least in pigs, is produced exclusively from leucine
supplement used by many exercise enthusiasts and more re- (Van Koevering and Nissen 1992). Plasma concentrations of
cently used in a medically related nutritional product to reduce HMB normally range from 1 to 4 ␮mol/L, but can increase 5-
wasting of muscle tissue in AIDS (Clark et al. 2000, Nissen et to 10-fold after leucine is fed (Nissen and Abumrad 1997).
al. 1996a, 1996b and 1997, Nissen and Abumrad 1997). The The cytosolic dioxygenase enzyme has been characterized ex-
major benefit of HMB appears to be a reduction in muscle tensively and differs from the mitochondrial KIC-dehydroge-
damage and/or reduced protein catabolism, which results in nase enzyme in that the dioxygenase enzyme is a cytosolic
improved gains in muscle size and strength when combined enzyme, whereas the dehydrogenase enzyme is found exclu-
with exercise (Nissen et al. 1996b). Although several animal sively in the mitochondrion (Sabourin and Bieber 1981 and
studies have shown that HMB consumption does not cause 1983). Recently, the KIC-dioxygenase enzyme was found to be
adverse effects (Nissen et al. 1994a and 1994b, Peterson et al. identical to the tyrosine dioxygenase enzyme (Janskiewiez et
1999a and 1999b, Van Koevering et al. 1993 and 1994), until al. 1996). It has been calculated that, under normal condi-
now there has not been a comprehensive analysis of safety data tions, ⬃5% of leucine oxidation proceeds via this pathway
collected on HMB-fed humans. (Van Koevering and Nissen 1992).
HMB is a metabolite of the amino acid leucine and is Numerous biochemical studies have shown that HMB is a
produced endogenously in both animals and humans. The first
precursor of cholesterol (Bloch 1944, Rabinowitz 1955, Rud-
step in the metabolism of leucine is transamination to ␣-ke-
ney 1957, Zabin and Bloch 1951). HMB in the cytosol of liver
toisocaproate (KIC). HMB is then produced from KIC by the
cytosolic enzyme KIC-dioxygenase (Sabourin and Bieber and muscle is first converted to cytosolic ␤-hydroxy-␤-meth-
ylglutarate-Co-A (HMG-CoA), which can then be used for
cholesterol synthesis (Rudney 1957). Thus HMB can serve as
1
a precursor for cellular cholesterol synthesis especially in tis-
To whom correspondence and reprint requests should be addressed.
2
Abbreviations used: BUN, blood urea nitrogen; CPK, creatine phosphoki- sues such as muscle that rely on de novo synthesis of choles-
nase; GGT, ␥-glutamyl transpeptidase; HMB, ␤-hydroxy-␤-methylbutyrate; HMG- terol. The working theory for HMB action is that stressed or
Co-A, ␤-hydroxy-␤-methylglutarate-Co-A; KIC, ␣-ketoisocaproate; LDH, lactate damaged muscle cells may not be able to make sufficient
dehydrogenase; MCH, mean corpuscular hemoglobin; MCHC mean corpuscular
hemoglobin concentration; MCV, mean corpuscular volume; SGOT, aspartate HMG-CoA to support adequate cholesterol synthesis for cell
aminotransferase; SGPT, alanine aminotransferase. functions, including proper functioning of cell membranes.

0022-3166/00 $3.00 © 2000 American Society for Nutritional Sciences.

Manuscript received 14 January 2000. Initial review completed 15 February 2000. Revision accepted 3 April 2000.

1937

Downloaded from [Link]

by guest
on 06 February 2018
 1938 NISSEN ET AL.

TABLE 1
Subject demographics and study descriptors

Study number

1 2 3 4 5 6 7 8 9

Gender Male Male Female Male Male/Female Female Male/Female Male/Female Male
Type of exercise Weight Weight None Weight Running Weight Weight lifting Weight lifting Weight
lifting lifting lifting lifting lifting
Age range, y 19–30 18–22 20–41 18–38 21–47 19–47 63–81 62–79 18–29
Placebo, n 13 15 19 18 6 18 18 16 11
HMB,1 n 15 10 18 21 8 18 18 13 7
Study length, wk 3 7 4 4 5 4 8 8 8
Site2 ISU ISU ISU ISU ISU ISU ISU WSU BSU

1 Abbreviation: HMB, ␤-hydroxy-␤-methylbutyrate.

2 ISU represents Iowa State University, Ames, IA. WSU represents Wichita State University, Wichita, KS. BSU represents Ball State University,
Muncie, IN.

Therefore, supplemental HMB could be a convenient source phosphatase, ␥-glutamyl transpeptidase (GGT), aspartate amino-
of HMG-CoA in these cells to maintain adequate cholesterol transferase (SGOT) and alanine aminotransferase (SGPT) by kinetic
synthesis and, in turn, plasma membrane function. This con- analysis; calcium, phosphorous, iron, bilirubin, protein and albumin
tention is supported by the observation that supplementation by colorimetric methods; glucose, uric acid, blood urea nitrogen, total
of HMB can markedly decrease muscle damage as evidenced cholesterol, HDL cholesterol and triglycerides by enzymatic methods.
LDL and VLDL were then calculated using the Friedewald formula
by leaking of creatine phosphokinase (CPK) out of muscle (Friedewald et al. 1972) (Labcorp, St. Louis, MO). The whole-blood
cells (Cheng et al. 1998, Nissen et al. 1996b, Nissen and samples were also processed on the day of collection by Labcorp, and
Abumrad 1997). Also supporting this concept are several analyzed for hemoglobin, hematocrit, mean corpuscular volume
studies showing that inhibition of cholesterol synthesis in (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular
muscle with drugs can result in muscle damage (Pierno et al. hemoglobin concentration (MCHC), platelets, and red and white
1995), poor function (Bastiaanse et al. 1997, Yeagle 1991) and blood cell numbers using an automated cell counter (Coulter STKS,
even muscle cell death (Mutoh et al. 1999). Coulter, Fullerton, CA).
Supplemental HMB is usually taken by humans at a dosage Emotional profile. Weekly emotional profiles were measured by
of ⬃3 g/d. Therefore, the objective of this study is to report the administration of the Circumplex test of emotion (Russell 1980).
safety-related data collected on HMB given in a dose of 3 g/d This test consisted of 48 words that describe various emotions. Each
over a series of nine experiments. These experiments encom- word was scored by the subject from 1 to 5 with the degree of feeling
that the word evoked: 1-very slightly or not at all; 2-a little; 3-mod-
passed the young, the old, men and women, exercising and erately; 4-quite a bit; and 5-extremely. Groups of 6 words that
nonexercising subjects, and were from 3 to 8 wk in duration. characterized an emotional category were then summed into the
In each of these studies, three batteries of tests were used to following categories: 1) High Activation (aroused, astonished, stim-
determine safety. First, comprehensive blood work was com- ulated, surprised, active and intense); 2) Activated Pleasant Affect
pleted at regular intervals during each study. Second, the (enthusiastic, elated, excited, euphoric, lively and peppy); 3) Unac-
Circumplex test of emotion (Russell 1980) was given period- tivated Pleasant Affect (relaxed, content, at rest, calm, serene and at
ically during each study; and third, an adverse events ques- ease); 4) Pleasant Affect (happy, delighted, glad, cheerful, warm-
tionnaire was filled out at intervals during each study. To- hearted and pleased); 5) Low Activation (quiet, tranquil, still, inac-
gether, these measurements should indicate the safety and tive, idle and passive); 6) Unactivated Unpleasant Affect (dull,
tolerance of HMB in the general population. tranquil, still, inactive, idle and passive); 7) Unpleasant (unhappy,
miserable, sad, grouchy, gloomy and blue); 8) Activated Unpleasant
Affect (distressed, annoyed, fearful, nervous, jittery and anxious). A
SUBJECTS AND METHODS range of values between 6 and 30 were possible for each category.
Subjects. Each study was approved by a local Institutional Re- Adverse events. Weekly adverse events were measured by filling
view Board and signed informed consents were obtained from the out a questionnaire. This questionnaire examined common com-
subjects before participation in the respective study. Each study was plaints related to major organ systems. Subjects were asked if they
conducted as a double-blind, placebo-controlled study. Exclusion experienced any of these symptoms over the last 3 d. In most studies,
criteria were ongoing chronic disease processes, smoking, illegal drug the questionnaires were given weekly.
use, pregnancy and certain age and sex requirements, depending on General study protocols. Table 1 summarizes the basic subject
the study. characteristics and protocols used during the individual studies. Stud-
Blood sampling. Blood samples were collected from a superficial ies are referred to by study numbers one through nine. All of the
forearm vein into Vacutainers (Becton Dickinson, Vacutainer Sys- studies were conducted as randomized double-blind, placebo-con-
tems, Rutherford, NJ) after an overnight fast by the subjects. Both trolled studies in which neither the researchers nor the subjects knew
serum and whole blood were collected. A plasma sample was also which treatment contained the HMB. The subjects were informed as
collected, frozen and stored at ⫺20°C until analyzed for HMB con- to the purpose of each study in the informed consent statement. In all
centration by gas chromatography-mass spectrometry (Nissen et al. but Study 3, the purpose of the study was to assess the effects of HMB
1990b). The serum samples were processed the day of collection and on exercising subjects. In Study 3, the purpose was only to collect
analyzed for CPK by kinetic analysis (Integra 700 Analyzer, Roche safety data and no exercise was included. In each study, HMB was
Diagnostics, Indianapolis, IN). The following were determined using administered as the calcium salt of HMB, Ca(C5H10O3)2 䡠 H2O. This
an automated blood chemistry analyzer (Olympus 5231, Olympus, compound contained ⬃13% calcium; when consumed at 3 g/d, this
Melville, NY): sodium, potassium and chloride electrolytes (ion se- amounted to ⬃400 mg/d of additional calcium in the HMB-supple-
lective electrode); lactate dehydrogenase, plasma creatinine, alkaline mented subjects. In Studies 1–3, HMB was administered as the

Downloaded from [Link]

by guest
on 06 February 2018
 SAFETY OF ␤-HYDROXY-␤-METHYLBUTYRATE 1939

calcium salt mixed in either orange juice or a protein shake. In a supervised, resistance-training regimen 3 times/wk during the 4-wk
Studies 4 –9, capsules containing the HMB or placebo (rice flour) study. Blood samples were obtained from fasting subjects at the
were used. Potassium phosphate (KH2PO4) was added to the HMB beginning and again at the end of the study. Circumplex and adverse
supplement capsules to partially maintain the Ca/P ratio. This events questionnaires were given weekly during the study.
amounted to ⬃135 mg of P and ⬃170 mg of K per day. The RDA for Study 5. This study was conducted at Iowa State University with
Ca, P and K are 1000, 1000 and 3500 mg/d, respectively (NRC 1989). male and female runners. Subjects were paired according to their best
The placebo treatments were not balanced for any of these minerals. two-mile run time and past running experience and then were as-
Compliance with the treatment regimens was assessed by the use signed randomly to either a placebo (0 g HMB/d) or HMB (3 g
of plasma HMB levels and by written questionnaires. In all studies, HMB/d) treatment administered in capsules as described for Study 4.
samples of plasma were analyzed for HMB, which increases several Neither the subjects nor the researchers knew which capsules con-
fold after HMB is taken (Nissen and Abumrad 1997). In addition, tained HMB. Subjects continued to train and were instructed to take
written questionnaires were given at the end of Studies 1 and 2, four capsules, either placebo or HMB, 3 times/d with their meals. The
asking the subjects to report their compliance in taking the treat- supplements were taken for 5 wk and all subjects were given a log to
ments. In Study 3, subjects were given weekly questionnaires and record the day and the times they took their supplement. These logs
asked to report any noncompliance to the treatments; in Studies 5 were used to verify compliance in taking the supplement. Blood
and 7, the subjects were asked to record each time a dosage of either samples were obtained from fasting subjects at wk 0 and 5 of the
the placebo or HMB treatment was taken. In the other studies, study. The Circumplex survey and the adverse events questionnaire
compliance was assessed by plasma HMB only. were given weekly.
Study 1. This study was conducted at Iowa State University. Study 6. This 4-wk study was conducted at Iowa State University
Subjects were assigned randomly to either a placebo (0 g HMB/d) or and was similar to Study 4 except that women subjects were tested.
HMB (3.0 g HMB/d) treatment. Neither the subjects nor the re- As in Study 4, similar criteria were used to classify the subjects as
searchers knew which treatment contained HMB; two equal dosages either trained or untrained. Subjects were then randomly adminis-
per day in 0.47 L of orange juice (one 16-oz bottle) were adminis- tered either a placebo (0 g HMB/d) or HMB (3 g HMB/d). Treatment
tered. The subjects were instructed to drink one bottle of juice in the was as described for Study 4 and neither the subjects nor the research-
morning and another bottle of juice in the evening. The subjects then ers knew which capsules contained HMB. Fasting blood samples were
underwent a supervised, resistance-exercise program 3 times/wk for 3 obtained at wk 0, 2 and 4 of the study. The Circumplex survey and
wk and alternated exercising either the upper or lower body during the adverse events questionnaires were given on d 0, 7, 15, 20, 23 and
each exercise session. A blood sample was drawn from fasting subjects 28 of the study.
on Thursday and Friday of each week. The Circumplex survey and Study 7. This study was conducted at Iowa State University in
the adverse events questionnaire were given weekly. elderly subjects who were assigned randomly to either a placebo (0 g
Study 2. This study was conducted at Iowa State University in HMB/d) or HMB (3 g HMB/d) treatment for 8 wk. Neither subjects
healthy male athletes who were subjected to stringent exercise sched- nor researchers knew which capsules contained HMB; treatments
ules over the 7-wk study. Subjects were assigned randomly to either were administered as described for Study 4. All subjects underwent
a placebo or HMB treatment. The HMB treatment consisted of 3 strength training 3 times/wk for the duration of the study. Blood
servings/d of a protein drink mix, each containing 1 g of HMB for a samples were obtained from fasting subjects at wk 0, 4 and 8 of the
total of 3 g HMB/d. Similarly, the placebo treatment consisted of 3 study. The Circumplex and the adverse events questionnaires were
servings of an isocaloric carbohydrate drink mix without HMB. again given weekly.
Subjects were instructed to drink one serving with each meal and Study 8. This study was similar to Study 7 but was conducted at
neither the subjects nor the researchers knew which drink contained Wichita State University (Wichita, KS). Elderly subjects were as-
the HMB supplement. Subjects exercised a total of 6 d/wk with signed randomly to either a placebo (0 g HMB/d) or HMB (3 g
resistance exercise on 5 d/wk. Blood samples were obtained from HMB/d) treatment for 8 wk. Neither subjects now researchers knew
fasting subjects on d 0, 17, 31 and 49 of the study. The Circumplex which capsules contained HMB; treatments were administered as
survey and the adverse events questionnaire were given on the blood described for Study 4. Subjects participating in this study had no
sampling day. contraindications to exercise and had their physicians approval to
Study 3. This 4-wk study was conducted at Iowa State University participate. Before the initiation of the study, subjects had no expe-
and was designed to document the safety of HMB when consumed by rience with resistance training. Subjects underwent strength training
women. The subjects enrolled into the study were instructed to 2 times/wk on nonconsecutive days. On the three other days during
maintain any exercise program they were currently involved in and the week, subjects reported to an indoor track for walking and
not to start any new exercise programs during the study. Subjects were stretching, which consisted of 10 min of warm-up stretching, 40 min
assigned randomly to receive either a placebo (0 g HMB/d) or HMB of self-paced walking and then another 10 min of stretching for a
(3 g HMB/d) treatment. The placebo and HMB were administered cool-down period. Blood samples were obtained from fasting subjects
twice daily in equal dosages mixed in ⬃180 mL of orange juice. at wk 0, 2, 4, 6 and 8 of the study. The Circumplex and the adverse
Neither the subjects nor the researchers knew which drink contained events questionnaires were given weekly.
HMB. The subjects were instructed to consume one dosage of the Study 9. This 8-wk study was conducted at Ball State University
juice in the morning and one dosage in the evening. Blood samples (Muncie, IN) with untrained college-aged males. Subjects were
were taken after an overnight fast on d 0, 5, 10, 25 and 28. Subjects matched on the basis of body weight and then assigned randomly to
also filled out Circumplex and adverse event questionnaires on the receive either a placebo [0 mg HMB/(kg body weight 䡠 d)] or HMB
blood sampling day. [38 mg HMB/(kg body weight 䡠 d)] treatment. These dosages were
Study 4. In this study, conducted at Iowa State University, the chosen because they are equivalent to either 0 or 3 g HMB/d in a
effect of prior training on the effects of HMB supplementation was person of average body weight. Subjects and researchers did not know
examined in young males. Study participants were classified as trained which treatment contained HMB. Subjects underwent a resistance
if they currently participated in a regular resistance exercise program training program 3 times/wk during the study. Blood samples were
at least 3 times/wk. Subjects were classified as untrained if they had taken at wk 0, 1, 2, 4 and 8 after an overnight fast. The Circumplex
not participated in any regular weight lifting for at least 6 mo before and the adverse events questionnaires were again given weekly.
the study. Subjects were assigned randomly to receive either a placebo Statistical analysis. For blood, hematology and Circumplex
(0 g HMB/d) or HMB (3 g HMB/d) treatment administered in data, for which multiple samples or questionnaires were taken over
capsules. Again, neither the subjects nor the researchers knew which the course of the study, the net change in variables was estimated by
capsules contained HMB. The placebo capsules (0 g HMB/d) con- regressing the measured values vs. week of study and calculating the
tained rice flour, whereas each HMB capsule contained 250 mg of net change over the study for each subject. Because day-to-day
calcium HMB and 50 mg of KH2PO4. The subjects were instructed to variations in measurement occur, this linear estimate of change better
take either four placebo or four HMB capsules 3 times/d with meals, reflects the overall change due to treatment. For the Circumplex
which supplied either 0 or 3 g HMB/d. All subjects then underwent data, the differences (in slopes) were analyzed by using a Student’s t

Downloaded from [Link]

by guest
on 06 February 2018
 1940 NISSEN ET AL.

TABLE 2
Effect of ␤-hydroxy-␤-methylbutyrate (HMB) on the emotional profile of humans in a summary of 9 studies1

Net change
Emotional category Descriptive words/category Screen Delta due to HMB

High activation Aroused, astonished, stimulated, Control 12.3 ⫺1.34 ⫺0.25

surprised, active, intense HMB 12.1 ⫺1.59
Activated pleasant Enthusiastic, elated, excited, Control 14.9 ⫺2.20 0.11
affect euphoric, lively, peppy HMB 14.2 ⫺2.10
Unactivated Relaxed, content, at rest, calm, Control 18.8 ⫺3.28 ⫺0.36
pleasant affect serene, at ease HMB 18.5 ⫺3.64
Pleasant affect Happy, delighted, glad, cheerful, Control 18.5 ⫺2.83 0.28
warmhearted, pleased HMB 17.5 ⫺2.55
Low activation Quiet, tranquil, still, inactive, idle, Control 13.5 ⫺1.72 ⫺0.62
passive HMB 14.0 ⫺2.34
Unactivated, Dull, tired, drowsy, sluggish, Control 10.3 ⫺0.23 ⫺1.04*
unpleasant affect bored, droopy HMB 11.0 ⫺1.27
Unpleasant Unhappy, miserable, sad, Control 7.5 0.18 0.05
grouchy, gloomy, blue HMB 7.7 0.23
Activated Distressed, annoyed, fearful, Control 8.7 0.61 ⫺0.12
unpleasant affect nervous, jittery, anxious HMB 8.9 0.74

1 Results presented are means of the screening and change over the studies to the Circumplex emotional profile (n ⫽ 127 placebo and n ⫽ 129
HMB). Each word of the profile was scored from 1 (very slightly or not at all) to 5 (extremely) according to how the subject felt towards that word.
Category scores were then calculated by adding the scores of the 6 words in each category (6 –30 points possible in each category).
* Indicates a significant effect (P ⬍ 0.05) across all experiments as determined by a Student’s t test.

test. For the blood and hematology data, the changes were analyzed the pretreatment or treatment period categories. A trend was
using ANOVA (SAS, Cary, NC) with the main effects of experi- indicated for a decrease in the incidence of loss of appetite (P
ment, treatment and experiment ⫻ treatment interaction in the ⬍ 0.07); however, this difference was ⬍ 1.5%. A trend was
model. The adverse events data were analyzed as categorical data also indicated for a difference in stiff joints reported (P
(Agresti 1990). For statistical analysis, the subjects were categorized
by before (yes or no) and after treatment response (any yes ⫽ yes, all ⬍ 0.08). However, because the statistical model also took into
no ⫽ no) and statistical significance between treatments determined consideration starting values, the indication was that a differ-
using the Cochran-Mantel-Haenszel test (Agresti 1990). In the case ence in incidence in the starting populations existed rather
of Study 8, “before” treatment questionnaires were not given; there- than a treatment effect.
fore, the subjects were divided into groups of either a “Yes” or “No” Blood chemistry, hematology and pressure. Blood lipid
answer during the treatment period and statistical significance was profiles of all subjects and those subjects with total cholesterol
determined by Fisher’s Exact test (Agresti 1990). Statistical signifi- levels either ⬎5.17 or ⬍5.17 mmol/L (200 mg/dL) are shown
cance was determined for P ⱕ 0.05. A trend was determined to in Table 4. No significant differences existed in any of the
be 0.05 ⱕ P ⱕ 0.10 and no statistical significance was determined starting values except for triglycerides in the subjects whose
for P ⬎ 0.10.
total cholesterol was ⬎ 5.17 mmol/L. Significant decreases of
3.7% (P ⬍ 0.03) for total cholesterol and 5.7% (P ⬍ 0.05) for
RESULTS LDL cholesterol were observed in the subjects consuming 3 g
Compliance. Supplementation compliance was checked HMB/d. Because of the health risk of high levels of choles-
by measuring plasma levels of HMB after supplementation. In terol, the subjects were subclassified into two groups, those
addition, in Studies 1, 2, 3, 5 and 7, a self-reported record of subjects having a total cholesterol level ⬍ 5.17 mmol/L (200
compliance was also obtained from the subjects. Basal plasma mg/dL) and those subjects having a total cholesterol level
levels of HMB were between 1 and 4 ␮mol/L and rose to ⬎ 5.17 mmol/L. The average starting values for these groups
between 17 and 25 ␮mol/L after an overnight fast in subjects were 4.27 mmol/L in the low cholesterol group and 5.72
consuming 3 g CaHMB/d (Nissen and Abumrad 1997). Al- mmol/L in the high cholesterol group. Although significant
though this is indicative of compliance only immediately decreases in total cholesterol were observed in both groups,
before the sampling time, this method does give some indica- the difference was only 2.5% (P ⬍ 0.04) in the low cholesterol
tion of overall compliance during the studies. On the basis of group, whereas in the high cholesterol group, there was a 5.8%
both the written responses from the subjects and plasma HMB decrease (P ⬍ 0.03) or ⫺0.334 mmol/L (12.9 mg/dL). The
levels measures, no subjects were dropped from the analyses for decrease in total cholesterol can be attributed to a decrease in
noncompliance. LDL cholesterol, which declined by 4.2% (P ⬍ 0.09) and 7.3%
Circumplex emotional profile. The Circumplex Model of (P ⬍ 0.01) in the low and high cholesterol groups, respec-
Affect questionnaire was administered in all of the studies, and tively.
the summary of the results is shown in Table 2. The only HDL cholesterol did not change in the HMB-fed subjects
significant change measured indicated a decrease (P ⬍ 0.05) in but increased ⬃4% in the placebo group (P ⬍ 0.04 placebo vs.
the category “Unactivated Unpleasant Affect,” which is de- HMB). The high risk subgroup had no significant change in
scribed by the words dull, tired, drowsy, sluggish, bored and HDL cholesterol.
droopy. Resting blood pressures were measured in seven of the nine
Adverse effects. None of the questions asked had re- studies and the combined results are also shown in Table 4. In
sponses that led to a significant treatment effect (Table 3). the studies, supplementation with HMB resulted in a signifi-
Analysis of the data showed no significant differences in either cant decrease in systolic blood pressure of 4.4 mm Hg (P

Downloaded from [Link]

by guest
on 06 February 2018
 SAFETY OF ␤-HYDROXY-␤-METHYLBUTYRATE 1941

TABLE 3
Summary of adverse events questionnaires given during 9 studies in which subjects consumed
3 g/d of ␤-hydroxy-␤-methylbutyrate (HMB)1

Control HMB

Initial Treatment period Initial Treatment period Treatment

Characteristic incidence, % incidence, % incidence, % incidence, % effect2

Stomachache 2.7 3.7 4.6 3.3 0.54

Nausea/Vomit 1.4 2.9 1.3 2.3 0.54
Dizziness 1.4 6.0 1.3 4.1 0.42
Coughing 12.8 8.3 13.9 11.0 0.66
Wheezing 2.0 1.6 4.0 5.2 0.51
Chest pain 0.7 1.0 2.0 1.1 0.47
Weakness 4.7 4.4 4.0 4.1 0.27
Increased headache 10.1 6.4 9.3 8.0 0.71
Negative mood3 7.0 2.4 1.8 3.6 0.38
Rash 1.4 2.1 2.0 0.8 0.33
Dry scalp/Hair3 4.0 3.5 2.8 6.4 0.67
Dry skin 4.7 3.6 6.0 8.5 0.24
Nail changes 0.7 0.7 0.7 1.5 0.50
Ear pain 0.7 1.5 1.3 1.6 0.46
Decreased memory3 1.0 1.3 0.9 1.3 0.36
Itching4 3.3 2.8 3.2 4.3 0.63
Swelling4 0.8 1.7 1.6 2.1 0.82
Diarrhea4 5.8 4.7 4.8 2.4 0.15
Stiff joints4 11.7 12.4 4.0 9.3 0.08
Nose bleeds4 0.8 0.7 1.6 0.9 0.81
Heart burn4 3.3 4.1 4.0 2.2 0.85
Numbness4 0.8 2.2 0.0 1.9 0.42
Nasal congestion4 20.8 14.8 10.3 14.9 0.26
Ringing in ears4 4.2 7.7 6.3 9.1 0.80
Increased stress3 11.9 6.9 3.7 4.9 0.19
Decreased libido5 0.0 1.7 0.0 0.6 0.93
Constipation4 1.7 2.2 2.4 3.5 0.88
Shortness of breath4 2.5 1.7 2.4 3.9 0.79
Loss of appetite4 3.3 3.4 4.0 1.9 0.07
Loss of energy4 7.5 5.6 4.8 5.8 0.61
Blood in urine4 0.8 0.2 0.8 0.2 1.00
Blood in stool4 0.0 0.2 0.0 0.2 0.98

1 Means of the percentage of reported incidents during all studies in which the question was asked (n ⫽ 136 placebo and n ⫽ 133 HMB).
Subjects were asked to fill out the questionnaires on a weekly basis. All studies have initial starting questionnaire values except for Study 8 in which
case only the treatment period incidence is included in the means.
2 Cochran-Mantel-Haenszel statistic for probable differences between the treatment groups.
3 Question asked only in Studies 4 through 9 (n ⫽ 89 placebo and n ⫽ 91 HMB).
4 Question asked only in Studies 3 through 9 (n ⫽ 108 placebo and n ⫽ 108 HMB).
5 Question asked in all but Study 3 (n ⫽ 117 placebo and n ⫽ 116 HMB).

⬍ 0.05 placebo vs. HMB). When subjects were stratified into phorous, protein, albumin and globulin. A small but signifi-
high and low risk, the subjects with a systolic pressure of ⬎ 130 cant decrease (1.9%, P ⬍ 0.003) in potassium, the primary
mm Hg had an even greater decrease in systolic pressure, but intracellular electrolyte, was seen. Additionally, there was a
the net difference between the HMB and placebo groups 5% increase (P ⬍ 0.03) in the albumin/globulin ratio in the
remained at 4.9 mm Hg. Diastolic blood pressures were not HMB group. A significant experiment ⫻ treatment effect was
affected by HMB supplementation. also seen for the albumin/globulin ratio because a response did
Blood indicators of liver function are summarized in Table not occur in all experiments. Similarly, significant experiment
5. There were no significant differences between the placebo ⫻ treatment effects were observed for BUN, BUN/creatinine,
and HMB groups for any of the measured indicators [bilirubin, protein and globulin, which resulted from the fact that the
alkaline phosphatase, lactate dehydrogenase (LDH), SGOT, responses between the placebo and treatment groups differed
SGPT, GGT, and iron]. There were significant experiment by from experiment to experiment with no particular pattern.
treatment interactions for LDH, GGT, and iron. In the case of The only significant effect of HMB on blood hematology
LDH, this effect was likely due to the variation in response to was a 0.48% relative decrease (P ⬍ 0.05) in hematocrit across
the exercise component in the different studies because plasma all studies (Table 7). The initial values and final values for
LDH can also be derived from damaged muscle. hematocrit, however, were not significantly different between
General blood chemistries are shown in Table 6. No sig- the placebo and HMB groups. Final values for hematocrit were
nificant differences between the placebo and HMB groups 0.43 for both placebo and HMB groups. There were no signif-
were seen for CPK, glucose, uric acid, blood urea nitrogen icant experiment ⫻ treatment interactions for any of the
(BUN), creatinine, BUN/creatinine, sodium, chloride, phos- hematology variables measured.

Downloaded from [Link]

by guest
on 06 February 2018
 1942 NISSEN ET AL.

TABLE 4
Effect of 3 g/d supplemental ␤-hydroxy-␤-methylbutyrate (HMB) on blood lipid profiles and blood pressure (BP)
in a summary in humans in 9 clinical studies1
All subjects Lower risk subjects Higher risk subjects

Change HMB Exp ⫻ Change HMB Exp ⫻ Change HMB Exp ⫻

Screen Delta % effect2 HMB2 Screen Delta % effect2 HMB2 Screen Delta % effect2 HMB2

Subjects with cholesterol ⬍ 5.17 mmol/L Subjects with cholesterol ⱖ 5.17 mmol/L

Total cholesterol,
mmol/L
Control 4.68 ⫺0.034 ⫺0.7 4.24 0.057 1.3 5.72 ⫺0.036 ⫺0.6
HMB 4.76 ⫺0.176 ⫺3.7 0.03 0.92 4.29 ⫺0.106 ⫺2.5 0.04 0.46 5.72 ⫺0.334 ⫺5.8 0.03 0.21
HDL cholesterol,
mmol/L
Control 1.14 0.049 4.3 1.09 0.052 4.7 1.26 0.028 2.3
HMB 1.17 0.008 0.7 0.04 0.57 1.13 0.003 0.2 0.06 0.77 1.30 0.003 0.2 0.53 0.59
VLDL cholesterol,
mmol/L
Control 0.53 ⫺0.009 ⫺1.7 0.50 0.011 2.3 0.64 ⫺0.070 ⫺10.9
HMB 0.54 ⫺0.016 ⫺3.0 0.79 0.36 0.51 0.009 1.7 0.93 0.30 0.58 ⫺0.054 ⫺9.4 0.78 0.09
LDL cholesterol,
mmol/L
Control 2.97 ⫺0.062 ⫺2.1 2.66 ⫺0.008 ⫺0.3 3.83 0.034 0.9
HMB 3.05 ⫺0.173 ⫺5.7 0.05 1.00 2.69 ⫺0.114 ⫺4.2 0.09 0.60 3.90 ⫺0.284 ⫺7.3 0.01 0.33
Total/HDL
cholesterol
Control 4.31 ⫺0.16 ⫺3.7 4.09 ⫺0.12 ⫺2.9 4.96 ⫺0.12 ⫺2.4
HMB 4.24 ⫺0.17 ⫺4.0 0.95 0.49 3.96 ⫺0.07 ⫺1.8 0.63 0.32 4.64 ⫺0.31 ⫺6.7 0.22 0.42
Triglycerides,
mmol/L
Control 1.22 ⫺0.019 ⫺1.6 1.13 0.034 3.0 1.64 ⫺0.207 ⫺12.6
HMB 1.23 ⫺0.066 ⫺5.3 0.46 0.08 1.16 ⫺0.028 ⫺2.4 0.46 0.19 1.29 ⫺0.114 ⫺8.8 0.41 0.10
Subjects with systolic BP ⬍ 130 Subjects with systolic BP ⱖ 130

Systolic BP, mm
Hg
Control 133 ⫺0.8 ⫺0.6 121 2.0 1.7 143 ⫺4.4 ⫺3.1
HMB 134 ⫺4.4 ⫺3.3 0.05 0.16 121 0.9 0.7 0.76 0.18 144 ⫺9.3 ⫺6.5 0.05 0.39
Diastolic BP, mm
Hg
Control 73.0 ⫺3.6 ⫺4.9 68.7 ⫺2.5 ⫺3.6 77.7 ⫺5.4 ⫺6.9
HMB 73.7 ⫺3.7 ⫺5.0 0.96 0.52 67.9 ⫺3.3 ⫺4.9 0.70 0.42 78.3 ⫺5.0 ⫺6.4 0.83 0.30

1 Results presented are least-square means from the combined analysis of the nine studies (n ⫽ 134 placebo and n ⫽ 128 HMB for blood lipids
and n ⫽ 112 placebo and n ⫽ 110 HMB for blood pressure). Screen values are the initial values of the participating subjects and delta values are
the change over the study. Blood pressures were not measured in Studies 5 and 9.
2 P-values for HMB effect and Experiment by HMB interaction as determined by ANOVA.

DISCUSSION whose average starting cholesterol was ⬎ 5.17 mmol/L (200

mg/dL) is examined, the decrease was even greater, and HMB
Measurement of the safety of nutritional products in hu-
resulted in a 5.8 and 7.3% decrease in total and LDL choles-
mans is usually accomplished using surrogate markers of health
and pathology. In the studies outlined here, three general areas terol, respectively. This amounted to a drop of ⬃13 points
of well-being were assessed. Blood chemistry, hematology and (mg/dL) in total and a drop of 11 points (mg/dL) in LDL
blood pressure were measured as indicators of tissue and organ cholesterol. In contrast, subjects with cholesterol values
functions. Emotional profiles were measured as an indication ⬍ 5.17 mmol/L (200 mg/dL) did not have a significant de-
of subtle changes in metabolism/chemistry that might be man- crease in LDL cholesterol when HMB was given. This suggests
ifested in emotional changes. Last, negative changes in per- that HMB is more effective in lowering cholesterol when
ception of body functions were assessed by looking for possible blood levels are elevated above the current standard associated
symptoms. Together, these data across both men and women, with increased risk of heart disease.
in young and old and in exercising and sedentary people give There was also a difference in HDL cholesterol response
a good indication of the safety of HMB in the general human with HMB supplementation compared with placebo although
population. this difference appeared to disappear when total cholesterol
Physiologic and biochemical measurements indicated that was elevated. The increase in HDL cholesterol in the placebo
HMB affects primarily cholesterol metabolism. Across all ex- group is likely related to the exercise-induced rise in HDL
periments, decreases in total cholesterol and LDL cholesterol cholesterol seen in previous studies (Halle et al. 1999).
of 3.7 and 5.7%, respectively, were seen. If a subset of subjects A second cardiovascular risk factor affected by HMB was

Downloaded from [Link]

by guest
on 06 February 2018
 SAFETY OF ␤-HYDROXY-␤-METHYLBUTYRATE 1943

TABLE 5 resulted in minor changes in ionic balance (although this

represents only ⬃40, 14 and 5% of the estimated daily intake
Effect of 3 g/d supplemental ␤-hydroxy-␤-methylbutyrate of Ca, P and K, respectively) (NRC 1989).
(HMB) on blood indicators of liver function in humans The only other significant difference in blood chemistry was a
in a summary of 9 clinical studies1 significantly greater albumin/globulin ratio in HMB-supple-
mented subjects. This was primarily the result of a relative
% HMB Exp ⫻ increase in albumin and a relative decrease in globulin.
Screen Delta Change effect2 HMB2

Bilirubin, ␮mol/L
Control 12.7 ⫺1.7 ⫺13.5 TABLE 6
HMB 12.1 ⫺1.9 ⫺15.5 0.76 0.18
Alkaline Effect of 3 g/d supplemental ␤-hydroxy-␤-methylbutyrate
phosphatase, (HMB) on blood chemistry profiles in humans in a summary of
IU/L
Control 80 0.9 1.1 9 clinical studies1
HMB 83 0.5 0.6 0.81 0.76
Lactate % HMB Experiment
dehydrogenase, Screen Delta Change effect2 by HMB2
IU/L
Control 184 ⫺8.9 ⫺4.8 Creatine kinase, U/L
HMB 176 ⫺5.0 ⫺2.8 0.44 0.03 Control 157 84 54
SGOT,3 IU/L HMB 141 59 42 0.43 0.13
Control 35.8 ⫺4.7 ⫺13.1 Glucose, mmol/L
HMB 31.6 ⫺2.8 ⫺8.9 0.42 0.56 Control 4.88 0.089 1.8
SGPT, IU/L HMB 4.70 0.222 4.7 0.12 0.15
Control 21.3 0.15 0.7 Uric acid, mmol/L
HMB 20.2 0.37 1.8 0.83 0.66 Control 0.263 0.0077 2.9
GGT, IU/L HMB 0.259 0.0131 5.1 0.39 0.37
Control 19.9 ⫺1.39 ⫺7.0 Blood urea nitrogen,
HMB 23.1 ⫺1.88 ⫺8.1 0.57 0.04 mmol/L
Iron, mmol/L Control 2.46 0.023 0.9
Control 0.0197 ⫺0.002 ⫺9.7 HMB 2.50 ⫺0.002 ⫺0.1 0.75 0.02
HMB 0.0201 ⫺0.003 ⫺16.3 0.16 0.008 Creatinine, ␮mol/L
Control 91.1 ⫺0.80 ⫺0.9
1 Results presented are least-square means from the combined HMB 91.1 ⫺2.48 ⫺2.7 0.15 0.26
analysis of the nine studies (n ⫽ 134 placebo and n ⫽ 128 HMB). Blood urea nitrogen/
Screen values are the initial values of the participating subjects and creatinine, wt/
delta values are the change over the study. wt
2 P-values for HMB effect and Experiment by HMB interaction as Control 14.3 0.27 1.9
determined by ANOVA. HMB 14.4 0.27 1.9 0.99 0.006
3 Abbreviations used: SGOT, aspartate aminotransferase; SGPT, Sodium, mmol/L
alanine aminotransferase; GGT, ␥-glutamyl transpeptidase. Control 142 ⫺0.478 ⫺0.3
HMB 142 ⫺0.783 ⫺0.6 0.16 0.20
Potassium, mmol/L
Control 4.27 0.036 0.8
blood pressure. Subjects given HMB had a decrease of ⬃3% in HMB 4.30 ⫺0.082 ⫺1.9 0.003 0.07
systolic blood pressure. Diastolic blood pressure decreased sim- Chloride, mmol/L
ilarly to the decrease seen in the placebo-supplemented group. Control 105 0.20 0.2
HMB 105 0.51 0.5 0.62 0.92
This change in blood pressure may be attributed to the addi- Calcium, mmol/L
tional calcium intake of ⬃400 mg/d with the Ca-HMB sup- Control 2.40 ⫺0.032 ⫺1.3
plement. There is a variety of experimental evidence showing HMB 2.42 ⫺0.045 ⫺1.9 0.30 0.21
that calcium supplementation can affect blood pressure (Bir- Phosphorous,
kett 1998, Kok et al. 1986, McCarron and Reusser 1999, mmol/L
Sempos et al. 1986). Meta-analysis of the data in these exper- Control 1.22 0.006 0.5
HMB 1.26 ⫺0.013 ⫺1.0 0.45 0.16
iments indicates that calcium lowers systolic blood pressure at Protein, g/L
the rate of ⬃0.39 mm Hg/(100 mg Ca䡠d). On the basis of this Control 72.6 ⫺1.6 ⫺2.2
equation, the 400 mg of Ca given in the current study should HMB 72.3 ⫺1.7 ⫺2.4 0.92 0.04
have lowered systolic blood pressure ⬃1.56 mm Hg. The Albumin, g/L
measured effect in subjects with systolic blood pressure ⬎130 Control 43.9 ⫺0.6 ⫺1.4
was much greater, i.e., 9.3 mm Hg total or 4.9 mm Hg net, HMB 43.6 ⫺0.2 ⫺0.5 0.20 0.82
Globulin, g/L
suggesting that either the CaHMB is a more biologically Control 28.7 ⫺1.1 ⫺3.8
available source of calcium (Sousa et al. 1996) or that HMB is HMB 28.7 ⫺1.5 ⫺5.2 0.20 0.005
having a more direct effect on cardiovascular function. Albumin/Globulin
Blood potassium decreased ⬍ 2% in the HMB-supple- Control 1.52 0.028 1.8
mented subjects. The reason for this small decrease is not HMB 1.50 0.075 5.0 0.03 0.02
known but is likely due to the slightly different mineral intakes
1 Results presented are least-square means from the combined
between the placebo- and the HMB-supplemented groups.
analysis of the nine studies (n ⫽ 134 placebo and n ⫽ 128 HMB).
The placebo supplement was not balanced for the calcium, Screen values are the initial values of the participating subjects and
phosphorus or potassium, which were contained in the HMB delta values are the change over the study.
treatments. Thus, the daily dose of ⬃400 mg of calcium, ⬃135 2 P-values for HMB effect and Experiment by HMB interaction as
mg of phosphorus and ⬃170 mg of potassium could have determined by ANOVA.

Downloaded from [Link]

by guest
on 06 February 2018
 1944 NISSEN ET AL.

TABLE 7 The only significant difference in blood hematology was an

⬃0.5% lower hematocrit in the HMB-supplemented group.
Effect of 3 g/d supplemental ␤-hydroxy-␤-methylbutyrate This difference was also reflected nonsignificantly in ⬃1%
(HMB) on blood hematology in humans in a summary of 9 lower red cell numbers and hemoglobin. An explanation for
clinical studies1 these differences is not apparent, but in all cases, the ending
values were not significantly different between the treatments.
% HMB Experiment Other experiments in animals indicate that HMB can increase
Screen Delta Change effect2 by HMB2 hematocrit, decrease granulocytes and increase lymphocytes
(Sandberg et al. 1997, Nissen et al. 1994c, Siwicki et al.
WBC,3 ⫻109/L
Control 6.56 ⫺0.25 ⫺3.8 1998b).
HMB 6.55 ⫺0.54 ⫺8.2 0.11 0.73 Previous research has shown that feeding HMB to a wide
RBC, ⫻1012/L variety of animals has not had any adverse effects on health
Control 4.76 ⫺0.08 ⫺1.7 or growth (Nissen et al. 1994a and 1994b, Peterson et al.
HMB 4.78 ⫺0.14 ⫺2.9 0.08 0.13 1999a and 1999b, Van Koevering et al. 1993 and 1994). In
Hemoglobin, g/L fact, the most consistent effect of HMB in animals has been
Control 146 ⫺2.5 ⫺1.7
HMB 147 ⫺3.8 ⫺2.6 0.12 0.41 a positive effect on health. In sheep, pigs, cattle, chickens
Hematocrit, L/L and fish, there have been positive effects noted on both
Control 0.432 ⫺0.0039 ⫺0.9 immune function and improved resistance to infectious
HMB 0.434 ⫺0.0087 ⫺2.0 0.05 0.10 agents (Nissen et al. 1990a, 1994b and 1994c, Ostaszewski
MCV, ␮m3 et al. 1998, Peterson et al. 1999a and 1999b, Siwicki et al.
Control 91 0.7 0.8 1998a and 1998b).
HMB 91 0.4 0.4 0.12 0.30
MCH, pg The Circumplex Model of emotion has been used in hu-
Control 30.8 ⫺0.14 ⫺0.5 mans to assess emotional balance. In this model, emotion is
HMB 30.8 ⫺0.10 ⫺0.3 0.87 0.52 described as a continuum (circle) of positive (pleasant) and
MCHC, g/L negative (unpleasant) emotions that are either activated or
Control 344 ⫺2.6 ⫺0.8 unactivated. By scoring how the subject feels about a word (an
HMB 339 ⫺2.1 ⫺0.6 0.62 0.02
Platelets, ⫻109/L
emotional feeling), a profile of emotional balance is estimated.
Control 237 ⫺2.0 ⫺0.8 In the current study, the only significant change in emotion
HMB 236 ⫺5.2 ⫺2.2 0.47 0.60 was a decrease in “Unactivated Unpleasant Affect” when
Polycytes, % HMB was supplemented. This emotional category is described
Control 54.0 ⫺0.38 ⫺0.7 by the words dull, tired, drowsy, sluggish, bored and droopy.
HMB 53.2 ⫺0.85 ⫺1.6 0.64 0.57 Subjects taking HMB appeared to have less of this negative
Lymphocytes, %
Control 35.6 ⫺0.53 ⫺1.5
emotion. This study did not attempt to assess the cause of this
HMB 36.3 ⫺0.01 0 0.53 0.47 effect, but most of the studies did involve some sort of exercise
Monocytes, % component; it may be that HMB enhanced the effects of this
Control 6.29 0.88 14.0 exercise, which could translate to less of an Unactivated
HMB 6.12 1.19 19.4 0.24 0.41 Unpleasant Affect.
Eosinophils, % Incidences of all adverse symptoms were quite low except
Control 3.15 0.08 2.5 for coughing and nasal congestion, which indicates many
HMB 3.39 ⫺0.21 ⫺6.2 0.24 0.28
Basophils, % subjects had upper respiratory infections during the study.
Control 0.82 0.09 11.0 Although the responses to questions did not differ between the
HMB 0.85 0.01 1.2 0.41 0.43 placebo and the HMB groups, there were some trends for
Polycytes, ⫻109/L differences. The HMB-supplemented subjects tended to have
Control 3.59 ⫺0.12 ⫺3.3 less diarrhea and less loss of appetite. There was also a trend for
HMB 3.55 ⫺0.36 ⫺10.1 0.15 0.87 a greater number of stiff joints reported in the HMB-supple-
Lymphocytes,
⫻109/L mented group. However, the HMB-supplemented group
Control 2.30 ⫺0.14 ⫺6.1 started out with a much lower incidence of stiff joints and the
HMB 2.33 ⫺0.17 ⫺7.3 0.61 0.14 ending value was actually numerically lower than the placebo-
Monocytes, ⫻109/L supplemented group, suggesting this was simply a chance oc-
Control 0.41 0.04 9.8 currence.
HMB 0.40 0.04 10.0 0.79 0.41 In summary, the only definitive effects of HMB were posi-
Eosinophils, ⫻109/L
Control 0.21 ⫺0.007 ⫺3.3 tive in nature, especially relating to lowering plasma choles-
HMB 0.22 ⫺0.028 ⫺12.7 0.19 0.52 terol and blood pressure. These data suggest that the popular
Basophils, ⫻109/L use of supplemental HMB at 3 g/d as an ergogenic aid for
Control 0.065 0.006 9.2 exercise is well tolerated and safe in humans.
HMB 0.062 0.007 11.3 0.95 0.92

1 Results presented are least-square means from the combined

ACKNOWLEDGMENTS
analysis of the nine studies (n ⫽ 133 placebo and n ⫽ 129 HMB). We gratefully acknowledge the contribution of John Rathmacher
Screen values are the initial values of the participating subjects and
in collecting the data for this manuscript. We also thank Philip
delta values are the change over the study.
2 P-values for HMB effect and Experiment by HMB interaction as Dixon for consultation on statistical analysis of these data.
determined by ANOVA.
3 Abbreviations used: WBC, white blood cells; MCV, mean corpus-
cular volume; MCH, mean corpuscular hemoglobin; MCHC mean cor-
LITERATURE CITED
puscular hemoglobin concentration. Agresti, A. (1990) Categorical Data Analysis. John Wiley and Sons, New York, NY.
Bastiaanse, E. M., Hold, K. M. & Van der Laarse, A. (1997) The effect of

Downloaded from [Link]

by guest
on 06 February 2018
 SAFETY OF ␤-HYDROXY-␤-METHYLBUTYRATE 1945

membrane cholesterol content on ion transport processes in plasma mem- Nissen, S., Van Koevering, M. & Webb, D. (1990b) Analysis of ␤-hydroxy-␤-
branes. Cardiovasc. Res. 33: 272–283. methyl butyrate in plasma by gas chromatography and mass spectrometry.
Birkett, N. J. (1998) Comments on a meta-analysis of the relation between Anal. Biochem. 188: 17–19.
dietary calcium intake and blood pressure. Am. J. Epidemiol. 148: 223–228. Ostaszewski, P., Kozlowska, E., Siwicki, A. K., Krzyzanowski, J., Fuller, J. C., Jr.
Bloch, K. (1944) Some aspects of the metabolism of leucine and valine. J. Biol. & Nissen, S. (1998) The immunomodulating activity of dietary 3-hydroxy-
Chem. 115: 255–263. 3-methylbutyrate (HMB) in weanling pigs. J. Anim. Sci. 76 (suppl. 1): 136.
Cheng, W., Phillips, B. & Abumrad, N. (1998) Effect of HMB on fuel utilization, Peterson, A. L., Qureshi, M. A., Ferket, P. R. & Fuller, J. C., Jr. (1999a) En-
membrane stability and creatine kinase content of cultured muscle cells. hancement of cellular and humoral immunity in young broilers by the dietary
FASEB J. 12: A950 (abs.). supplementation of ␤-hydroxy-␤-methylbutyrate. Immunopharm. Immuno-
Clark, R. H., Feleke, M., Din, M., Yasmin, T., Singh, G. Khan, F. & Rathmacher, toxicol. 21: 307–330.
J. A. (2000) Nutritional treatment for acquired immunodeficiency virus- Peterson, A. L., Qureshi, M. A., Ferket, P. R. & Fuller, J. C., Jr. (1999b) In vitro
associated wasting using ␤-hydroxy-␤-methylbutyrate, glutamine and argi- exposure with ␤-hydroxy-␤-methylbutyrate enhances chicken macrophage
nine: a randomized, double-blind, placebo-controlled study. JPEN 24: 133– growth and function. Vet. Immunol. Immunopathol. 67: 67–78.
139.
Pierno, S., De Luca, A., Tricarico, D., Roselli, A., Natuzzi, F., Ferrannini, E., Laico,
Friedewald, W. T., Levy, R. I. & Fredrickson, D. S. (1972) Estimation of the
M. & Camerino, D. C. (1995) Potential risk of myopathy by HMG-CoA
concentration of low-density lipoprotein cholesterol in plasma, without use of
reductase inhibitors: a comparison of pravastatin and simvastatin effects on
the preparative ultracentrifuge. Clin. Chem. 18: 499 –502.
Halle, M., Berg, A., Baumstark, M. W. & Keul, J. (1999) Association of physical membrane electrical properties of rat skeletal muscle fibers. J. Pharmacol.
fitness with LDL and HDL subfractions in young healthy men. Int. J. Sports Exp. Ther. 275: 1490 –1496.
Med. 20: 464 – 469. Rabinowitz, J. L. (1955) The biosynthesis of radioactive ␤-hydroxyisovaleric
Janskiewiez, J., Popov, K. M. & Harris, R. A. (1996) ␣-Ketoisocaproate dioxy- acid in rat liver. J. Am. Chem. Soc. 77: 1295–1297.
genase, a dual-specificity enzyme corresponds to alloantigen F. FASEB J. 9: Rudney, H. (1957) The biosynthesis of beta-hydroxy-beta-methylglutaric acid.
A318 (abs.). J. Biol. Chem. 227: 363–377.
Kok, F. J., Vandenbroucke, J. P., Van Der Heide-Wessel, C. & Van Der Heide, Russell, J. A. (1980) A circumplex model of affect. J. Pers. Soc. Psychol. 39:
R. M. (1986) Dietary sodium, calcium, and potassium, and blood pressure. 1161–1178.
Am. J. Epidemiol. 123: 1043–1048. Sabourin, P. J. & Bieber, L. L. (1981) Subcellular distribution and partial
McCarron, D. A. & Reusser, M. E. (1999) Finding consensus in the dietary characterization of an ␣-ketoisocaproate oxidase of rat liver: formation of
calcium-blood pressure debate. J. Am. Coll. Nutr. 18: 398S– 405S. ␤-hydroxyisovaleric acid. Arch. Biochem. Biophys. 206: 132–144.
Mutoh, T., Kumano, T., Nakagawa, H. & Kuriyama, M. (1999) Role of tyrosine Sabourin, P. J. & Bieber, L. L. (1983) Formation of ␤-hydroxyisovalerate by an
phosphorylation of phospholipase C gamma1 in the signaling pathway of ␣-ketoisocaproate oxygenase in human liver. Metabolism 32: 160 –164.
HMG-CoA reductase inhibitor-induced cell death of L6 myoblasts. FEBS Lett. Sandberg, L., Miller, P. & Fuller, J. C., Jr. (1997) Effect of ␤-hydroxy-␤-
446: 91–94. methylbutyrate on the physiological response to exercise and conditioning in
National Research Council (1989) Recommended Dietary Allowances, 10th horses. J. Anim. Sci. 75 (suppl. 1): 198.
ed. National Academy Press, Washington, DC. Sempos, C., Cooper, R., Kovar, M G., Johnson, C., Drizd, T. & Yetley, E. (1986)
Nissen, S. L. & Abumrad, N. N. (1997) Nutritional role of the leucine metabolite Dietary calcium and blood pressure in National Health and Nutrition Exami-
␤-hydroxy-␤-methylbutyrate (HMB). J. Nutr. Biochem. 8: 300 –311. nation Surveys I and II. Hypertension 8: 1067–1074.
Nissen, S., Faidley, T. D., Zimmerman, D. R., Izard, R. & Fisher, C. T. (1994a) Siwicki, A. K., Morand, M., Fuller, J. C., Jr., Nissen, S., Kazua, K. I., Glabski, E. &
Colostral milk fat percentage and pig performance are enhanced by feeding Ostaszewski, P. (1998a) Influence of 3-hydroxy-3-methylbutyrate (HMB)
the leucine metabolite ␤-hydroxy ␤-methylbutyrate to sows. J. Anim. Sci. 72: on specific cellular immune response after in vitro and in vivo immunization
2332–2337. with Yersinia ruckeri antigen. J. Anim. Sci. 76 (suppl. 1): 136.
Nissen, S., Fuller, J. C., Jr., Sell, J., Ferket, P. R. & Rives, D. V. (1994b) The Siwicki, A. K., Ostaszewski, P., Fuller, J. C., Jr., Nissen, S. & Studnicka, M.
effect of ␤-hydroxy-␤-methylbutyrate on growth, mortality and carcass qual- (1998b) In vitro effects of 3-hydroxy-3-methylbutyrate (HMB) on measures
ities of broiler chickens. Poult. Sci. 73: 137–155. of immune function and immunocompetence in fish. J. Anim. Sci. 76 (suppl.
Nissen, S., Kuhlman, G. & Roth, J. A. (1990a) Modulation of ovine lymphocyte
1): 136.
function by leucine and leucine metabolites. In: Amino Acids: Chemistry,
Sousa, M. F., Abumrad, N. N., Martins, C., Nissen, S. & Riella, M. C. (1996)
Biology and Medicine (Lubec, G., ed.), pp. 817– 820. ESCOM, Leiden, The
Calcium ␤-hydroxy-␤-methylbutyrate. Potential role as a phosphate binder in
Netherlands.
uremia: in vitro study. Nephron 72: 391–394.
Nissen, S., Morrical, D. & Fuller, J. C., Jr. (1994c) The effects of the leucine
catabolite ␤-hydroxy-␤-methylbutyrate (HMB) on the growth and health of Van Koevering, M. T., Gill, D. R., Smith, R. A., Owens, F. N., Nissen, S. & Ball, R. L.
growing lambs. J. Anim. Sci. 77 (suppl. 1): 243. (1993) Effect of ␤-hydroxy-␤-methyl butyrate on the health and performance
Nissen, S., Panton, L., Fuller, J., Jr., Rice, D. & Sharp, R. (1997) Effect of of shipping-stressed calves. Oklahoma State Univ. Res. Rep., pp. 312–316.
feeding ␤-hydroxy-␤-methylbutyrate (HMB) on body composition and Van Koevering, M. & Nissen, S. (1992) Oxidation of leucine and ␣-ketoisoca-
strength of women. FASEB J. 11: A150 (abs.). proate to ␤-hydroxy-␤-methylbutyrate in vivo. Am. J. Physiol. 262: E27–E31.
Nissen, S., Panton, L., Wilhelm, R. & Fuller, J. C., Jr. (1996a) Effect of Van Koevering, M. T., Dolezal, H. G., Gill, D. R., Owens, F. N., Strasia, C. A.,
␤-hydroxy-␤-methylbutyrate (HMB) supplementation on strength and body Buchanan, D. S., Lake, R. & Nissen, S. (1994) Effects of ␤-Hydroxy-␤-
composition of trained and untrained males undergoing intense resistance methylbutyrate on performance and carcass quality of feedlot steers. J. Anim.
training. FASEB J. 10: A287 (abs.). Sci. 72: 1927–1935.
Nissen, S., Sharp, R., Ray, M., Rathmacher, J. A., Rice, J., Fuller, J. C., Jr., Yeagle, P. L. (1991) Modulation of membrane function by cholesterol. Bio-
Connelly, A. S. & Abumrad, N. N. (1996b) The effect of the leucine me- chimie 73: 1303–1310.
tabolite ␤-hydroxy ␤-methylbutyrate on muscle metabolism during resis- Zabin, I. & Bloch, K. (1951) The utilization of butyric acid for the synthesis of
tance-exercise training. J. Appl. Physiol. 81: 2095–2104. cholesterol and fatty acids. J. Biol. Chem. 192: 261–266.

Downloaded from [Link]

by guest
on 06 February 2018