Beruflich Dokumente
Kultur Dokumente
Intermountain Heart Institute, Intermountain Medical Center, Salt Lake City, UT, USA
Genetic Epidemiology Division, Department of Medicine, University of Utah, Salt Lake City,
UT, USA
c
Cardiology Division, Department of Medicine, University of Utah, Salt Lake City, UT, USA
b
Received 29 April 2012; received in revised form 27 September 2012; accepted 30 September 2012
Available online 7 December 2012
KEYWORDS
Fasting;
Erythrocyte;
Hemoglobin;
Human growth
hormone;
Energy restriction;
Caloric restriction;
FGF-21
Abstract Background and aims: Routine, periodic fasting is associated with a lower prevalence of coronary artery disease (CAD). Animal studies show that fasting may increase longevity
and alter biological parameters related to longevity. We evaluated whether fasting initiates
acute changes in biomarker expression in humans that may impact short- and long-term health.
Methods and results: Apparently-healthy volunteers (N Z 30) without a recent history of fasting
were enrolled in a randomized cross-over trial. A one-day water-only fast was the intervention
and changes in biomarkers were the study endpoints. Bonferroni correction required
p 0.00167 for significance (p < 0.05 was a trend that was only suggestively significant). The
one-day fasting intervention acutely increased human growth hormone (p Z 1.1 104), hemoglobin (p Z 4.8 107), red blood cell count (p Z 2.5 106), hematocrit (p Z 3.0 106), total
cholesterol (p Z 5.8 10 5 ), and high-density lipoprotein cholesterol (p Z 0.0015), and
decreased triglycerides (p Z 1.3 10 4 ), bicarbonate (p Z 3.9 10 4 ), and weight
(p Z 1.0 107), compared to a day of usual eating. For those randomized to fast the first day
(n Z 16), most factors including human growth hormone and cholesterol returned to baseline
after the full 48 h, with the exception of weight (p Z 2.5 104) and (suggestively significant)
triglycerides (p Z 0.028).
Conclusion: Fasting induced acute changes in biomarkers of metabolic, cardiovascular, and
general health. The long-term consequences of these short-term changes are unknown but
* Corresponding author. Intermountain Heart Institute, 5121 S. Cottonwood St., Salt Lake City, UT 84107, USA. Tel.: 1 801 507 4708; fax:
1 801 507 4792.
E-mail address: benjamin.horne@imail.org (B.D. Horne).
1
For the Intermountain Heart Collaborative Study Group.
0939-4753/$ - see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.numecd.2012.09.007
1051
Introduction
Routine, periodic fasting was previously associated with
a lower prevalence of coronary artery disease (CAD) [1].
That potential fasting benefit was hypothesized by logical
deduction based on multiple lines of evidence: 1) lower
cardiovascular mortality among Utahns and specifically
members of the Church of Jesus Christ of Latter-day Saints
(LDS, or Mormons) [2,3], 2) lower CAD risk among
Utahnsdindependent of tobacco use [1], and 3) a review of
shared behaviors of the LDS population [1]. That study
utilized no a priori biological hypothesis and a chance
finding could not be ruled out, but the finding was subsequently replicated in an independent population [4]. Those
epidemiologic studies provided no mechanistic explanation, but an association of fasting with lower risk of diabetes was observed [1,4].
Parallel research in animal models provides some potential insights that independently suggest a health benefit from
fasting, perhaps via the glucose metabolism pathway [5,6].
Some evidence indicates that a human genetic locus related
to those findings in animals is associated with glucose
concentration [7]. While those data hint at the hypothesis of
a metabolic effect, whether short-term fasting causes
persistent changes that benefit human health is unknown.
This study evaluated whether a short-term fasting
intervention acutely alters biochemical and physiological
parameters related to long-term risk. The objective was to
test whether 1 day of water-only fasting changes blood
levels of biomarkers of metabolic, cardiovascular, and
general health among apparently-healthy individuals.
Methods
The FEELGOOD trial was a two-day clinical cross-over trial
of apparently-healthy individuals that used a randomized
Latin square design. Participants acted as their own control
and, thus, intervention and control were matched for
participant factors such as age, sex, and body mass index
(BMI). Participants were included if they had not fasted
>12 h per episode for >1 year prior to enrollment, not
routinely participated in caloric restriction (i.e., deliberate
limitation to <80% of FDA recommended caloric intake) for
>2 years, and not deliberately skipped meals as a routine
dietary practice.
Study population
Individuals aged 18e70 years of either gender and any
ethnicity were eligible for enrollment. Individuals with
previous myocardial infarction, peripheral vascular disease,
or stroke, and those with immunodeficiency diseases were
Study endpoints
Thirty-six primary endpoints were designated, including
insulin, glucose, creatinine, sodium, potassium, calcium,
chloride, bicarbonate, blood urea nitrogen (BUN), homeostasis model assessment-insulin resistance (HOMA-IR, using
the HOMA2 algorithm) [8], human growth hormone (HGH),
white blood cell count (WBC), red blood cell count (RBC),
hemoglobin, hematocrit, platelet count, red cell distribution
width (RDW), mean corpuscular hemoglobin (MCH), mean
corpuscular hemoglobin concentration (MCHC), mean
platelet volume (MPV), mean corpuscular volume (MCV), total
cholesterol, LDL-C, high-density lipoprotein cholesterol
(HDL-C), triglycerides, total cholesterol to HDL-C (TC:HDL)
ratio, high-sensitivity C-reactive protein (hsCRP), sitting and
supine SBP and DBP, weight, waist circumference, adiponectin (ratio of high molecular weight to total), glycogen-like
protein-1 (GLP-1), and fibroblast growth factor-21 (FGF-21).
The original design designated the primary endpoint as
changes in expression of the FOXA1, FOXA2, and FOXA3
genes (www.clinicaltrials.gov: NCT01059760), which
hypothesis was based on prior animal evidence of a fasting
effect [5,6]. Patient mRNA samples were collected by
venipuncture in Paxgene Blood RNA tubes and total plasma
mRNA was isolated using Paxgene Blood RNA kit (PreAnalytiX) by manufacturer recommendations with oncolumn treatment of RNaseeFree DNase 1 (Qiagen). Foxa
transcripts were quantified with TaqMan gene expression
assays (Applied BioSystems, Inc) and GAPDH was used as an
endogenous control for normalization. Due to the number
of samples with mRNA transcripts beyond the detection
limit of the assay, gene expression of the three FOXA genes
were not analyzed statistically.
Statistical considerations
Change scores for each parameter were compared between
fasting and fed days by paired T-test. Linear regression was
1052
Results
Among N Z 30 participants, age averaged 43.6 13.5 years
(range: 19.5e64.4 years) and 66.7% were female. Baseline
BMI averaged 27.8 1.3 kg/m2 and mean hemoglobin A1C
was 5.4% 0.3% (range: 4.7%e5.9%). Only hsCRP differed
between randomization groups (Table 1). Mean age was
36.1 14.9 years for males and 48.6 11.1 years for
females (p Z 0.022), with mean BMIs of 29.6 7.5 kg/m2
and 27.0 6.1 kg/m2 (p Z 0.34). Adverse trial events were
2 (7%) mild headaches and 1 (3%) migraine.
For metabolic outcomes (Table 2), the one-day fast
significantly increased HGH (p Z 1.1 104). Nonsignificant (p > 0.00167) changes were observed (Table 2)
in glucose (p Z 0.0069), insulin (p Z 0.022), HOMA-IR
(p Z 0.013), and GLP-1 (p Z 0.041). No changes were
found in adiponectin (p Z 0.21) or FGF-21 (p Z 0.23).
General health endpoints increased by fasting were
hemoglobin (4.8 107), RBC (2.5 106), and hematocrit
(3.0 106) (Table 2), while bicarbonate was decreased
(3.9 104). Sodium demonstrated a strong trend downward (p Z 0.0018). No changes were found for platelet
count, potassium, or various other parameters (Table 2).
Cardiac risk factor outcomes are shown in Table 2.
Weight (1.0 107) and triglycerides (1.3 104) were
Table 1
Characteristic
p-value
Age (years)
Sex (female)
BMI (kg/m2)
Waist circumference (cm)
Glucose (mg/dL)
Hemoglobin A1C
SBP, supine (mmHg)
DBP, supine (mmHg)
Total cholesterol (mg/dL)
LDL-C (mg/dL)
HDL-C (mg/dL)
Triglycerides (mg/dL)
hsCRP (mg/L)b
43.3 13.6
68.8%
28.4 7.1
87.7 17.4
81.4 15.1
5.5% 0.3%
115 14
68.2 8.6
192 27
106 31
54.4 12.1
132 46
2.80 (1.62)
44.0 14.0
64.3%
27.3 6.1
86.1 16.2
83.4 19.7
5.3% 0.3%
114 9
68.8 7.0
190 34
103 27
56.3 19.7
156 87
0.75 (0.71)
0.91
0.80
0.69
0.82
0.76
0.19
0.88
0.86
0.92
0.78
0.75
0.35
0.008
a
Randomization was 1:1 by Latin square to either 1 day of fasting followed by a day of eating (fasting day first: n Z 16) or to a day of
eating followed by the fasting day (fed day first: n Z 14).
b
Values of hsCRP are median (standard error).
1053
Characteristic
Baseline value
Change while
fasting
Change when
fed
p-value for
change scores
HGH (ng/mL)
Glucose (mg/dL)
Insulin (mU/L)
HOMA-IRa
GLP-1 (ng/mL)
Adiponectinb
FGF-21 (pg/mL)
WBC (/mL)
Hemoglobin (g/dL)
RBC (106/mL)
Hematocrit (%)
Platelet count (103/mL)
MCV (fL)
MCH (pg)
MCHC (g/dL)
RDW (%)
MPV (fL)
Bicarbonate (mmol/L)
Sodium (mmol/L)
Chloride (mmol/L)
BUN (mg/dL)
Calcium (mg/dL)
Potassium (mmol/L)
Creatinine (mg/dL)
Total cholesterol (mg/dL)
LDL-C (mg/dL)
HDL-C (mg/dL)
Triglycerides (mg/dL)
TC/HDL ratio
Weight (kg)
Waist circumference (cm)
SBP, supine (mmHg)c
DBP, supine (mmHg)c
hsCRP (mg/L)d
0.32 0.47
83.3 15.8
23.6 17.8
0.66 0.51
1.77 0.79
0.34 0.17
190.3 261.4
6.0 1.4
14.3 1.3
4.7 0.4
41.6 3.6
253.5 58.2
89.11 3.56
30.66 1.48
34.4 0.5
13.60 0.82
8.17 0.83
27.0 2.5
139.1 2.0
103.7 2.7
16.0 3.7
9.4 0.3
4.29 0.28
0.848 0.118
190.7 30.7
103.8 33.8
55.9 14.3
134.2 77.0
3.63 1.15
81.3 25.0
86.7 16.7
114.0 13.6
67.2 8.8
1.35 (0.96)
1.54 2.63
8.0 15.4
19.0 29.3
0.35 0.53
0.27 0.51
0.02 0.25
41.7 128.9
0.32 1.31
0.46 0.37
0.15 0.13
1.3 1.3
2.8 16.3
0.12 0.68
0.01 0.33
0.02 0.49
0.04 0.32
0.02 0.41
1.4 1.7
1.4 2.1
1.3 2.0
1.2 3.9
0.18 0.34
0.02 0.30
0.005 0.09
9.5 10.6
23.1 35.1
3.7 5.3
52.4 55.5
3.3 17.7
1.5 0.9
0.57 1.4
3.1 8.4
1.7 7.4
0.00 (0.44)
1.14
5.9
1.6
0.02
0.07
0.07
9.5
0.12
0.39
0.11
1.1
0.9
0.18
0.07
0.01
0.03
0.03
0.9
0.7
0.9
1.8
0.12
0.02
0.002
4.8
10.3
1.0
11.2
0.0
0.5
0.54
1.0
1.2
0.10
1.1 104*
0.0069
0.022
0.017
0.041
0.21
0.23
0.32
4.8 107*
2.5 106*
3.0 106*
0.72
0.79
0.60
0.84
0.95
0.88
3.9 104*
0.0018
0.0037
0.0065
0.0087
0.67
0.89
5.8 105*
0.0051
0.0015*
1.3 104*
0.32
1.0 107*
0.13
0.19
0.32
0.60
1.66
16.6
16.2
0.30
0.51
0.25
108.8
1.48
0.52
0.17
1.5
13.8
0.73
0.40
0.52
0.37
0.26
2.2
2.0
2.3
2.6
0.33
0.32
0.06
11.0
28.2
3.6
38.1
0.2
0.9
3.0
9.6
8.9
(0.24)
*Correction for multiple comparisons required p 0.00167 for significance (p < 0.05 is suggestively significant).
a
HOMA-IR used the HOMA2 model [8].
b
Adiponectin values are given for the ratio of high molecular weight adiponectin to the total adiponectin.
c
Sitting SBP and DBP were similar (data not shown).
d
Values of hsCRP are median (standard error).
while the average after fasting was slightly higher for males
(2.23 2.61 vs. females: 1.98 2.41 ng/mL).
The fasting:fed ratios were 1.05 and 1.03 for hemoglobin
among females and males (p Z 0.22), respectively (results
for RBC and hematocrit were similar), 0.93 and 0.98 for
bicarbonate (p Z 0.07), and 1.07 and 1.03 for total
cholesterol (p Z 0.025, HDL-C results were similar). The
only difference in the ratios based on BMI was for bicarbonate (0.90, 0.95 and 0.997 for tertiles 1, 2, and 3,
respectively, p Z 0.009) and no differences were seen
based on age.
Among the 16 individuals randomized to fasting on the
first study day, all factors except weight (p Z 2.5 104)
returned to baseline after the full 48 h of the study
(Table 3), although triglycerides had a trend (p Z 0.028)
toward not having returned to baseline. In contrast, RBC
Discussion
Fasting (complete caloric desistance) and caloric restriction (partial limitation of calorie intake) are similar
methods of energy restriction that may influence human
health [10e12]. Differences may exist in the biological
source of their health benefits [5,6], consistent with the
knowledge that various cardiovascular pathways exist
through which dietary factors may act [13]. Previously, in
a population in which many patients routinely engage in
fasting over the life span for religious purposes (for 24 h
1054
1055
Table 3 Among only those participants (n Z 16) who were randomized to fast the first day (and eat on the second day), an
evaluation of the normalization of factors that were significantly altered by fasting.
Characteristic
Baseline value
(hour 0)
HGH (ng/mL)
Hemoglobin (g/dL)
RBC (106/mL)
Hematocrit (%)
Bicarbonate (mmol/L)
Total cholesterol (mg/dL)
HDL-C (mg/dL)
Triglycerides (mg/dL)
Weight (kg)
0.73
14.2
4.65
41.6
27.3
192
54.4
132
82.5
1.43
1.2
0.32
3.4
1.5
27
12.1
46
26.8
1.77a
1.1a
0.33a
3.2a
1.7a
32a
12.1a
33a
26.5a
0.60
1.3
0.35
3.5
2.5
28
11.6
38
26.7
*Correction for multiple comparisons required p 0.00167 for significance (p < 0.05 is suggestively significant).
a
The fasting value was changed compared to baseline (p < 0.05).
1056
Conclusions
A clinical interventional trial by 1 day of water-only fasting
recorded acute improvements in biomarkers of metabolic
risk, including profound changes in measures of erythrocyte
concentration, cholesterol, and HGH. These short-term
changes were normalized to baseline levels within oneday after fasting, but the specific factors that changed
suggest the potential for long-term metabolic and cardiovascular benefits. Further evaluation is needed to determine whether those short-term changes produce persistent
health benefits. If so, fasting could potentially be used to
prevent diabetes and CAD among those in a pre-diabetic
state with intermediate fasting glucose (e.g., glucose
100e125 mg/dL) or with metabolic syndrome. The possibility that fasting is an intensive biologically-programmed
treatment that can produce long-term improvements in
health indicates that water-only fasting deserves long-term
prospective clinical evaluation [39].
Funding
This study was funded by a grant from the Intermountain
Research and Medical Foundation, Salt Lake City, UT. The
funding organization had no role in the design or conduct of
the study, the collection, management, analysis, or interpretation of the data, or the preparation, review, or
approval of the manuscript.
Conflicts of interest
The authors have nothing to disclose.
References
[1] Horne BD, May HT, Anderson JL, Kfoury AG, Bailey BM,
McClure BS, et al. Usefulness of routine periodic fasting to
lower risk of coronary artery disease among patients undergoing coronary angiography. Am J Cardiol 2008;102:814e9.
[2] Lyon JL, Wetzler HP, Gardner JW, Klauber MR, Williams RR.
Cardiovascular mortality in Mormons and non-Mormons in
Utah, 1969e1971. Am J Epidemiol 1978;108:357e66.
[3] Enstrom JE. Health practices and cancer mortality among
active California Mormons. J Natl Cancer Inst 1989;81:
1807e14.
[4] Horne BD, Muhlestein JB, May HT, Carlquist JF, Lappe
DL,
Bair TL, et al. Relation of routine, periodic fasting to risk of
diabetes mellitus and coronary artery disease among patients
undergoing coronary angiography. Am J Cardiol 2012;109:
1558e62.
[5] Zhang L, Rubins NE, Ahima RS, Greenbaum LE, Kaestner KH.
Foxa2 integrates the transcriptional response of the hepatocyte to fasting. Cell Metab 2005;2:141e8.
[6] Panowski SH, Wolff S, Aguilaniu H, Durieux J, Dillin A. PHA4/Foxa mediates diet-restriction-induced longevity of C. elegans. Nature 2007;447:550e6.
[7] Xing C, Cohen JC, Boerwinkle E. A weighted false discovery
rate control procedure reveals alleles at FOXA2 that influence
fasting glucose levels. Am J Hum Genet 2010;86:440e6.
[8] Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA
modeling. Diabetes Care 2004;27:1487e95.
[28]
[29]
[30]
[31]
[32]
[33]
1057
[34]
[35]
[36]
[37]
[38]
[39]