1750 4732 2 7 PDF
1750 4732 2 7 PDF
1750 4732 2 7 PDF
Address: 1University of Oklahoma Health Sciences Center, Department of Family Medicine, Tulsa, OK 74120, USA, 2Oklahoma State University
Center for Health Sciences Department of Osteopathic Manipulative Medicine, Tulsa, OK 74107, USA and 3Oklahoma State University Center for
Health Sciences Department of Pharmacology and Physiology, Tulsa, OK 74107, USA
Email: Charles E Henley - [email protected]; Douglas Ivins - [email protected]; Miriam Mills - [email protected];
Frances K Wen - [email protected]; Bruce A Benjamin* - [email protected]
* Corresponding author
Abstract
Background: The relationship between osteopathic manipulative treatment (OMT) and the autonomic nervous system
has long been acknowledged, but is poorly understood. In an effort to define this relationship, cervical myofascial release
was used as the OMT technique with heart rate variability (HRV) as a surrogate for autonomic activity. This study
quantifies that relationship and demonstrates a cause and effect.
Methods: Seventeen healthy subjects, nine males and eight females aged 19–50 years from the faculty, staff, and students
at Oklahoma State University Center for Health Sciences College of Osteopathic Medicine, acted as their own controls
and received interventions, administered in separate sessions at least 24 hours apart, of cervical myofascial OMT, touch-
only sham OMT, and no-touch control while at a 50-degree head-up tilt. Each group was dichotomized into extremes of
autonomic activity using a tilt table. Comparisons were made between measurements taken at tilt and those taken at
pre- and post-intervention in the horizontal.
The variance of the spectral components of HRV, expressed as frequencies, measured the response to change in position
of the subjects. Normalized low frequency (LF) and high frequency (HF) values, including LF/HF ratio, were calculated
and used to determine the effect of position change on HRV.
Results: Predominantly parasympathetic responses were observed with subjects in the horizontal position, while a 50-
degree tilt provided a significantly different measure of maximum sympathetic tone (p < 0.001). Heart rate changed in all
subjects with change in position; respirations remained constant. When OMT was performed in a sympathetic
environment (tilt), a vagal response was produced that was strong enough to overcome the sympathetic tone. There was
no HRV difference between sham and control in either the horizontal or tilt positions.
Conclusion: The vagal response produced by the myofascial release procedure in the maximally stimulated sympathetic
environment could only have come from the application of the OMT. This demonstrates the association between OMT
and the autonomic nervous system. The lack of significance between control and sham in all positions indicates that HRV
may be a useful method of developing sham controls in future studies of OMT.
Trial registration: clinicaltrials.gov NCT00516984.
Page 1 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
Page 2 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
or less than 90 mmHg, or failure of heart rate to increase design, the subjects acted as their own controls and partic-
with passive tilt (50-degrees head-up). Long-distance run- ipated in each of three separate sessions during which
ners and other conditioned athletes also were excluded. only one of the experimental conditions (control, OMT,
sham) was administered. The order of administering the
The study proposal was submitted and approved after full experimental conditions was partly randomized. Specifi-
review of both the Oklahoma State University College of cally, at the first session, each subject was randomly
Osteopathic Medicine (IRB#2000417) and the University assigned to one of the two treatment conditions, OMT or
Of Oklahoma College Of Medicine (IRB#12024) Institu- sham. Subjects were then administered the other treat-
tional Review Boards. ment condition during the second session. The control
condition was administered as the third condition. The
General study design subjects were not told which treatment they were given at
The general experimental design is shown in Figure 1. the time. Random assignment to treatment conditions
There were three experimental conditions: a) control was conducted to control for within subject variations
involving no intervention; b) OMT treatment involving that might affect physiological responses. Only data for
cervical myofascial release, and c) sham treatment involv- subjects who completed all three sessions were analyzed
ing the placement of hands in the cervical region. The pro- and reported here.
tocol lasted 30 minutes. Changes in body position
(horizontal vs 50-degree head-up tilt) were incorporated After first obtaining a physician-read normal resting ECG,
into the study to facilitate interpretation of autonomic subjects were randomly assigned to a treatment condition
tone measures. ECG data were recorded continuously to begin the study. Subjects completed all three experi-
throughout the experiment. Analysis of autonomic tone mental sessions with a minimum of 24 hours separating
was standardized and occurred in the last 5 minutes of each session. Each session was conducted at the same time
horizontal 1 period, the last 5 minutes of tilt, and during of day, between 1:30 p.m. and 4:00 p.m. The subjects were
the first 5 minutes of heart rate recovery in horizontal 2 instructed to eat a light lunch without caffeine, no later
period, as shown by the shaded sections of Figure 1. Treat- than 12:00 noon. When they entered the exam room,
ment, either OMT or sham, was administered for 2 min- ECG electrodes were attached and the subject placed on
utes and started 30 seconds after tilt started. the tilt table. During an initial equilibration period the
subjects were acclimated to the tilt table by moving them
Subjects who met inclusion criteria were enrolled in the from the horizontal to head-up position several times.
study and consent was obtained. In a repeated measures This usually took about 5 minutes. The change in body
Time (min) 0 10 20 30
Session
Control *
OMT OMT
Sham Sham
Figure 1
Experimental protocol
Experimental protocol. Shaded areas show the times when data were analyzed for each subject. *= No treatment
Page 3 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
position was completely passive for the subjects. They did culated as the ratio of the power in the considered fre-
not exert any effort or assist in the change of position. quency band to the sum of the HF and LF powers: LFnu =
100*LF/(HF+LF); HFnu = 100*HF/(HF+LF). The ratio of
After equilibration, the following protocol was followed LFnu/HFnu was used as an index of autonomic balance.
for the control session: 10 minutes at horizontal, 10 min-
utes at 50-degrees head-up tilt, then back to horizontal for Statistical analysis
10 minutes. No treatment was given during the control The study is a two-way (3 × 3) repeated measures design,
session. For the OMT and sham treatment sessions, the with each subject acting as their own control. Analysis of
cervical myofascial release or sham procedure was initi- the data was conducted using within subjects analysis of
ated 30 seconds after the subject was placed in the head- variance (ANOVA). Factors were three experimental con-
up position and continued for two minutes. Preliminary ditions (Control, OMT, Sham) and three positions (Hori-
studies (unpublished observations from our pilot study of zontal 1, Tilt, Horizontal 2). Dependent variables were
nine subjects) indicated that it took subjects an average of heart rate, normalized LF (LFnu), normalized HF (HFnu),
30 seconds to reach a new steady state in heart rate after the ratio of LFnu/HFnu (LF/HF ratio), and respiration rate.
repositioning to head-up. At the end of treatment phase Post-hoc comparisons were performed with Bonferroni
(control, OMT, or sham), the subjects remained in the adjustment for multiple comparisons. Pilot study data
upright position for approximately 8 more minutes, and suggested a large difference in frequencies between condi-
then were returned to the horizontal position for a final tions indicating that approximations of sphericity would
10 minute period. be met without using correction factors. All data were ana-
lyzed using SPSS version 12. A power analysis, based on
The cervical myofascial release technique used in this the pilot data, showed a need for an N of 14 using an
study was adopted from the monograph of Common alpha of 0.05 and a beta of 0.20 (1-beta of 0.80), with an
OMT Techniques by Kenneth Graham, Oklahoma State estimated 25% difference in mean ratios for the three con-
University Center for Health Sciences.[12] The sham pro- ditions and a standard deviation of 3.6.
cedure was accomplished by placing the researcher's
hands in the same position on the neck as would occur if Data handling and security
initiating the cervical myofascial release technique. The Data acquisition and analysis was accomplished on a lap-
position is maintained for the same amount of time as the top computer which was never connected to the Internet.
OMT; however no pressure is applied to the cervical area. It was password protected and stored under lock and key
when not in use. All subjects received a data ID number
Data collection that was used to code their records. The file linking subject
A three lead ECG was recorded from the subject and at the name to ID was destroyed at the end of the experiment.
same time a respiratory trace was recorded with a piezo
chest transducer RM-204 (iWorx; Dover, NH). Data were Results
collected by an analog-to-digital converter (iWorx HK Seventeen (61%) of 28 subjects completed all three
214; Dover, NH) with a sampling frequency of 200 Hz phases of the study and had clean HRV data to analyze.
and stored on a personal computer for later analysis. The 17 subjects that completed the study exceeded our
estimated need for an N of 14 by three subjects. There
Digital records were analyzed after the experiment. R-R were no adverse effects of the study. Analysis of the sub-
intervals were determined by an R-wave peak detection jects not completing the study showed that two of the 28
algorithm. Each data record was reviewed manually. Time subjects did not have an increase in heart rate in response
series analysis consisted of 300 seconds of data containing to head-up tilt. Nine of the 28 had scheduling conflicts or
consecutive intervals of sinus rhythm. Spectral analysis moved out of town before the data collection was com-
was carried out with HRV analysis software from the Bio- pleted. The 17 subjects completing the study consisted of
medical Signal Analysis Group, Department of Physics, 9 men and 8 women. Their average systolic blood pressure
University of Kuopio, Finland. The 300 second R-R inter- was 120 mmHg ± 2 and diastolic blood pressure was 71
vals were interpolated by cubic-spline and resampled at 2 mmHg ± 2. Table 1 shows mean and 95% confidence
Hz to obtain equidistant data samples. The power spec- intervals for heart rate, LFnu, HFnu, and LF/HF parameters
trum was computed by Fourier Transform. Total power for the three conditions at each position.
was defined in the range of 0.01–0.40 Hz. Low frequency
power was defined as 0.04–0.15 Hz and high frequency Figure 2 shows the effect of body position and treatment
power as 0.15–0.40 Hz. Normalization of LF and HF indi- on heart rate. Passively changing body position from hor-
ces was used to minimise the impact of a change of the izontal to the 50-degree head-up position resulted in a sig-
total power in the analysis of the balance between LF and nificant increase in heart rate for the control, sham, and
HF. Normalization of power in LF and HF ranges were cal- OMT conditions, F(2,15) = 94.82, p < 0.001. Although
Page 4 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
Table 1: Mean and 95% Confidence Intervals (CI) for Heart Rate, LFnu, HFnu, and LF/HF Parameters
Position Mean Lower Bound Upper Bound Mean Lower Bound Upper Bound Mean Lower Bound Upper Bound
Heart Rate H1 67.61 61.79 73.44 65.64 60.42 70.86 62.43 57.89 66.96
(bpm)
Tilt 80.51 74.52 86.51 78.73 72.03 85.43 71.74 66.29 77.20
H2 65.43 59.85 71.01 64.03 58.74 69.32 59.53 55.44 63.62
LFnu H1 53.21 44.05 62.36 44.97 35.63 54.31 40.08 28.80 51.35
Tilt 80.19 76.00 84.38 76.93 71.25 82.61 58.33 49.87 66.79
H2 53.30 43.90 62.70 47.98 39.32 56.64 41.49 30.48 52.50
HFnu H1 47.89 38.41 57.38 54.43 44.53 64.33 59.37 47.37 71.37
Tilt 20.15 15.73 24.57 21.57 16.43 26.71 41.44 32.40 50.48
H2 47.50 37.61 57.39 51.38 42.22 60.53 56.60 45.64 67.54
LF/HF H1 1.46 0.97 1.96 1.17 0.53 1.81 1.09 0.41 1.77
Tilt 4.85 3.62 6.08 4.44 2.92 5.96 1.83 1.11 2.56
H2 1.60 0.90 2.29 1.29 0.58 2.00 1.04 0.49 1.60
H1 = first horizontal position, Tilt = 50-degree head-up position, H2 = second (recovery) horizontal position. LFnu = normalized low frequency,
HFnu = normalized high frequency, LF/HF = low frequency to high frequency ratio. The protocol was completed three times, Control, Sham, and
OMT.
heart rate tended to be lower in the OMT condition, the levels when the subjects returned to the horizontal posi-
change in heart rate that occurred when the subjects tion.
assumed the head-up position was similar in all three con-
ditions. Heart rate for all conditions returned to baseline The effects of body position and treatment on heart rate
variability in the frequency domains are shown in Figure
3. Normalized power in the low frequency range (Figure 3
upper panel) significantly increased in the control and
90 sham conditions when subjects were moved from hori-
zontal to head-up tilt, F(2,15) = 47.35, p < 0.001. The
Heart Rate (beats/min)
* *
magnitude of change was similar in the control and sham
80
*# conditions. However, when OMT treatment was applied
in the head-up position, the increase in LFnu was signifi-
70 cantly less when compared to the increases in LFnu that
occurrerd in the control and sham conditions (p < 0.001).
60 LFnu returned to baseline in all three experimental condi-
tions when the subjects were returned to the horizontal
position.
50
Horizontal 1 Tilt Horizontal 2
In the head-up tilt position, normalized high frequency
Control OMT Sham power (Figure 3, middle panel) in the control and sham
conditions significantly decreased, F(2,14) = 43.005, p <
Figureof2body position and treatment on heart rate
Effect 0.001. The change in HFnu was not as great in the OMT
Effect of body position and treatment on heart rate. treatment condition (p < 0.001).
Values are mean ± SEM. * indicates a significant difference (p
< 0.001) within a condition between H1 and Tilt positions; # As shown in Figure 3 (lower panel), head-up tilt signifi-
indicates a significant difference (p < 0.001) between control cantly increased the LF/HF ratio in all three conditions,
and OMT and sham and OMT at Tilt position. F(2,15) = 30.49, p < 0.001, reflecting an increase in sym-
pathetic tone. The increase in LF/HF ratio was due to an
Page 5 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
100 20
* *
80 18
Respiratory Rate
*#
(breaths/min)
LF(nu)
60 16
40
14
20
12
0
Horizontal 1 Tilt Horizontal 2 10
Horizontal 1 Tilt Horizontal 2
100 Control OMT Sham Control OMT Sham
80
Figureof4body position and treatment on respiratory rate
Effect
HF(nu)
60
*#
40
Effect of body position and treatment on respiratory
* *
* rate. Values are mean ± SEM. There were no significant dif-
20 ferences within or between conditions.
0
Horizontal 1 Tilt Horizontal 2
6 * OMT * Sham Heart rate variability and spectral indices of heart rate var-
Control
5 iability can be influenced by respiratory rate. Respiration
was measured continuously throughout the study. As
LF/HF
4
3 shown in Figure 4, respiration did not change with body
*# position and was similar in all three conditions, F(2,15) =
2
1 0.65, p > 0.50.
0
Horizontal 1 Tilt Horizontal 2
Discussion
Although the osteopathic profession has long recognized
Control OMT Sham
a relationship between the autonomic nervous system
and the function of the body in health and disease, the
Figure
Effect
HF ratio
of3body position and treatment on LFnu, HFnu and LF/ ability to demonstrate a causal relationship between OMT
Effect of body position and treatment on LFnu, HFnu and the autonomic nervous system has been problematic.
and LF/HF ratio. Values are mean ± SEM. LF = Low Fre-
Understanding the ability of manipulative medicine to
quency, HF = High Frequency, nu = normalized. * indicates a
significant difference (p < 0.001) within a condition between influence sympathovagal balance and recognizing its lim-
H1 and Tilt positions; # indicates a significant difference (p < itations is critical in defining and refining OMT outcomes.
0.001) between control and OMT conditions and sham and The goal of the present study was to address this problem
OMT conditions in the Tilt position. by using noninvasive measures of sympathovagal bal-
ance. This study tested the hypothesis that cervical myo-
fascial release increased vagal tone. Using an experimental
protocol in which a tilt table was used to set a range for
sympathetic and vagal modulation of heart rate, the
increase in LFnu and a decrease in HFnu power. Cervical results of the study showed that the initiation of OMT in
myofascial release resulted in a lower LF/HF ratio in the a head-up tilt position resulted in a change back, nearly to
head-up tilt phase by decreasing LFnu and increasing HFnu baseline levels, to a more parasympathetic tone. This rep-
power in comparison to control and sham. The initiation resented a shift from a sympathetic to a parasympathetic
of OMT in head-up tilt resulted in a change back, nearly to environment, demonstrating the effect of the OMT treat-
baseline, to a more parasympathetic tone. The LF/HF ratio ment in overcoming the sympathetic tone. The results
returned toward baseline conditions when the subjects established a clear association between the effects of OMT
were returned to the horizontal position. and changes in autonomic activity.
The LF/HF ratios in the head-up tilt position for the con- Results of this study have demonstrated, for the first time,
trol condition and the sham condition were over 2.5 a quantitative relationship between osteopathic manipu-
times the LF/HF ratio in the OMT condition, as can be lation and sympathovagal balance. That is, cervical myo-
seen in Table 1. This is a statistically significant difference fascial release shifts sympathovagal balance from the
(p < 0.001) between the control and sham conditions at sympathetic to the parasympathetic nervous system. The
tilt and the OMT condition at tilt. present study simplifies the analysis and reduces nonspe-
Page 6 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
cific influences by putting the results of HRV in the con- power decreased from 47.9 to 20.2 and the LF/HF ratio
text of changes that occur during passive head-up tilt. In increased from 1.5 to 4.9 when the subjects were tilted.
this study, passive tilt was used to create a graded change The difference between the two studies is the intervention
in sympathovagal balance and it did this with minimal of OMT in the present study. Further inspection of Table
changes in central cardiovascular drive, muscle afferents 1 in the present study shows that, when cervical myofas-
nerve activity, respiration, and cortical activity.[13] The cial release was applied in the tilt position, as much as
effect in this study of overcoming a maximally stimulated 79% of the increase in the LF/HF ratio was abolished. This
sympathetic tone created by the tilt position means that convincingly shows an effect of OMT on sympathovagal
the observed parasympathetic response could only have balance.
come from the effect of the OMT. In effect, the OMT
results are compared against changes in spectral HRV This study shows the utility of using HRV as a method for
measures that occur in response to a physiological change measuring autonomic change. It also demonstrates the
in body position. positive association between OMT and the autonomic
nervous system as the HRV changes with the initiation of
The mechanism for the effect of cervical myofascial release OMT. This is the first time the association between OMT
on autonomic balance was not investigated in this study. and autonomic activity has been quantified in this way,
However, the role of neck afferents with projections to and establishes a plausible mechanism for how OMT
vestibular nuclei may be a possibility. The role of neck works in the body. The change in HRV seen with initiation
afferents and vestibulo-sympathetic reflexes in cardiovas- of OMT has physiological significance since it is being
cular control is receiving increasing attention. It is well compared to pre-calibrated positional changes. In this
documented that head down neck flexion can activate a study, OMT intervention is the only explanation for pro-
variety of autonomic reflexes that affect vascular control. ducing a significant vagal response in an otherwise strong
In this regard, Lee et al[14] studied the effect of head sympathetic environment. The implications of this study
down neck flexion on heart rate variability and found that are important in laying a foundation for understanding
this maneuver significantly affected autonomic balance as how OMT facilitates physiologic change, taking the use of
measured by the normalized LF/HF ratio. OMT out of the arena of empiricism. The ability to use
HRV as a surrogate for autonomic activity also highlights
Spectral measure of HRV and autonomic balance are sub- its use in sham controls where the concern has always
ject to some variability due to their sensitivity to changes been that the positioning of a subject or laying on of
in hemodynamic state, stress, and health. It is generally hands could somehow exert its own effect. In this study
recognized that respiration has an important effect on we observed that the sham and control conditions were
heart rate variability.[15] Respiration was continuously not significantly different from each other.
measured in this study and was not affected by body posi-
tion or condition (control, treatment, sham). Therefore, Conclusion
the results of the study were not influenced by alterations This study demonstrated quantitatively that cervical myo-
in respiration. fascial release shifts sympathovagal balance from the sym-
pathetic to the parsympathetic nervous system. This study
The changes observed in heart rate and spectral analysis in employed one particular type of OMT technique, cervical
response to passive head-up tilt in this study are compara- myofascial release. Future studies should be aimed at
ble to the change observed in other studies.[10,16-18] demonstrating the same effect on the autonomic nervous
Montano et al[16] used power spectrum analysis to assess system using other techniques, such as high velocity/low
the changes in sympathovagal balance during graded tilt. amplitude. The subjects used in the present study were
They found graded changes in normalized units of low free of pathology. Studies showing the effect on auto-
and high frequency power and the LF/HF ratio. The mag- nomic change when OMT is applied in a disease state are
nitude of changes in their study (60-degrees head-up tilt) needed for further understanding of the physiological
was very similar to the changes seen in our control sub- mechanism.
jects. Based on Figure 3 from Montano et al,[16] LFnu
increased from approximately 40 in the horizontal posi- List of Abbreviations
tion to 75 in the 60-degree head-up position. Normalized OMT: Osteopathic Manipulative Treatment; HRV: Heart
HF power decreased from approximately 50 to 20 and the rate variability; ECG: electrocardiogram; LF: Low fre-
LF/HF ratio increased from around 1 to 9 when the sub- quency; HF: High frequency; LFnu: Low frequency, nor-
jects were tilted. Inspection of Table 1 in the present study malized; HFnu: High frequency, normalized; LF/HF: Low
shows comparable changes for our control condition: frequency to high frequency ratio; IRB: Institutional
LFnu increased from 53.2 in the horizontal position to review board; ANOVA: Analysis of variance.
80.2 in the 50-degree head-up position. Normalized HF
Page 7 of 8
(page number not for citation purposes)
Osteopathic Medicine and Primary Care 2008, 2:7 http://www.om-pc.com/content/2/1/7
References
1. Anderson RE, Seniscal C: A comparison of selected Osteopathic
treatment and relaxation for tension-type headaches. Head-
ache 2006, 46:1273-1280.
2. Burns DK, Wells MR: Gross range of motion in the cervical
spine: The effects of Osteopathic muscle energy technique in
asymptomatic subjects. JAOA 2006, 106(3):137-142.
3. Celander E, Koenig AJ, Celander DR: Effect of Osteopathic
manipulative therapy on autonomic tone as evidenced by
blood pressure changes and activity of the fibrinolytic sys-
tem. JAOA 1968, 67(9):1037-1038.
4. Sergueef N, Nelson KE, Glonek T: The effect of cranial manipu-
lation on the Traube-Hering-Mayer Oscillation as measured
by Laser-Doppler Flowmetry. Altern Ther Health Med 2002,
8(6):74-76.
5. Kautzner J, Camm AJ: Clinical relevance of heart rate variabil-
ity. Clin Cardiol 1997, 20(2):162-168.
6. Sztajzel J: Heart rate variability: a noninvasive electrocardio- Publish with Bio Med Central and every
graphic method to measure the autonomic nervous system. scientist can read your work free of charge
Swiss Med Wkly 2004, 134(35-36):514-522.
7. Heart rate variability. Standards of measurement, physio- "BioMed Central will be the most significant development for
logical interpretation, and clinical use. Eur Heart J 1996, disseminating the results of biomedical researc h in our lifetime."
17(3):354-381. Sir Paul Nurse, Cancer Research UK
8. Nakamura Y, Yamamoto Y, Muraoka I: Autonomic control of
heart rate during physical exercise and fractal dimension of Your research papers will be:
heart rate variability. J Appl Physiol 1993, 74(2):875-881. available free of charge to the entire biomedical community
9. Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Pizzinelli P,
Sandrone G, Malfatto G, Dell'Orto S, Piccaluga E, Turiel M, Baselli G, peer reviewed and published immediately upon acceptance
Cerutti S, Malliani A: Power spectral analysis of heart rate and cited in PubMed and archived on PubMed Central
arterial pressure variabilities as a marker of sympatho-vagal
interaction in man and conscious dog. Circ Res 1986, yours — you keep the copyright
59(2):178-193.
Submit your manuscript here: BioMedcentral
http://www.biomedcentral.com/info/publishing_adv.asp
Page 8 of 8
(page number not for citation purposes)