Effect of a virtual reality-enhanced exercise protocol after coronary artery bypass grafting.

Research over the past decade has shown that virtual reality (VR)--enhanced programs can be effective in the treatment of patients recovering from brain injury or with psychological disorders. (1-4) Virtual reality technology also can be successfully incorporated into exercise programs. For instance, a patient on a stationary cycle ergometer and provided with a VR display and 3-dimensional (3D) stereo headphones can go for a virtual "ride in the hills" during an exercise session. To ensure that the patient is carrying out the prescribed exercise properly, performance can be monitored online. (5) Virtual reality--enhanced exercise programs have been shown to generate feelings of positive involvement, revitalization, and calmness in participants. (5) These feelings have enabled the training periods to become more intense and to last longer. (6) As our team previously reported, VR-based rehabilitation can assist senior citizens who are healthy in maintaining endurance, increasing their target exercise intensity, and enhancing their total energy consumption during exercise. (7) In that experiment, elderly subjects who were healthy exercised on friction-braked ergometers connected by computer to a flat screen depicting a country road with 2 cyclists riding on it. The flow on the screen was matched to the speed of the ergometer, and subjects were instructed to cycle at a pace equal to that of the riders on the screen as the experimenter adjusted the load on the ergometer according to the protocol and recorded the subject's blood pressure, heart rate (HR), and oxygen consumption (V[O.sub.2]).

Although better performance in subjects who were healthy was shown by that experiment, the effect of the VR environment on exercise capacity outcomes of long-term, simulation-based exercise training of people with cardiovascular disorders had not yet been investigated in any systematic studies. Cardiac rehabilitation programs are increasingly used by patients who have undergone coronary artery bypass grafting (CABG). (8) Over successive sessions, the patients' increasing physical endurance is matched by a similar increase in the training load used during these sessions. In an investigation of the effects of VR on patients who were receiving cardiac rehabilitation, Chuang et al (9) showed that the VR group achieved significantly higher values than the non-VR group for peak oxygen consumption (V[O.sub.2]peak), peak metabolic equivalents, and V[O.sub.2] level at which the anaerobic threshold was reached when these values were measured in follow-up maximum exercise tests. The patients in that investigation trained on a treadmill, either without the VR experience or with the treadmill linked by computer to a visual screen with a wide field of view, 3D auditory inputs and accelerator cards, and a graphic user interface allowing speeds and changes in the incline of the treadmill to be synchronized with changes in the scenery on the screen. Heart rate, blood pressure, and V[O.sub.2] were monitored in both groups as in the previous experiment with healthy subjects. After training twice weekly for 3 months on the treadmill, the VR group achieved better cardiac performance (as shown by a higher V[O.sub.2]peak) on the follow-up exercise tests than the control group. The present study, therefore, was designed to examine the effect of VR-enhanced programs for patients with CABG on the number of training sessions needed to reach target physiologic endpoints (which were, in this study, 85% maximum HR [HRmax] and 75% V[O.sub.2]peak).

Method

Participants

Each subject who participated in this study had received CABG between January and June 2004. Subjects were prospectively recruited from the cardiovascular surgery department at the Veterans Affairs Medical Center Taipei, Taipei, Taiwan, and were included if they qualified for the supervised outpatient cardiac rehabilitation programs (phase II). During this time, 24 subjects were recruited after undergoing bypass surgery. The placement of each subject in a treatment arm was decided by lottery. Two balls, designated A and B, were placed in a box. If the A ball was drawn, then the subject was assigned to group 1; that is, no VR would be used. If the B ball was drawn, the subject was assigned to group 2, and a VR experience would be provided during the rehabilitation sessions. Informed consent was obtained after the nature of the study procedures had been fully explained and understood.

Instrumentation

The Veterans Affairs Medical Center Taipei operates the Telepresence Cardiac Rehabilitation Program, in which users are physically active in and interactive with an imaginary 3D setting, as though they were physically in a real-life scenario. The system's graphic user interface permits speed alteration and treadmill incline adjustments in conjunction with scenery changes. The system also comprises a visual screen with a wide field of view, 3D auditory outputs, and 3D accelerator cards.

The Microsoft* Windows series operating systems form the principal operating environment for this model and include Windows 2000 Professional and XP. High-end PCs process and display the 3D simulations in real time through powerful PC engines, which are based on a 2.4-GHz Pentium IV processor with 512 MB of SDRAM and 3D accelerator cards. Within this protocol, the VR scenes show a Microsoft Direct 3D-constructed "virtual runner" model. For our program, the scenes were fused into standard 2-dimensional background descriptions. This process involved loading the virtual environment in stages, beginning with the basic scene description and followed by the 3D graphics, which were the nested descriptions.

The images were projected from behind the viewer through 3 projectors connected with computers. Three computers communicated with each other via a transmission control protocol or Internet transfer protocols, which made up a local area network. A virtual environment was displayed on three ~239-cm (94-in)-wide connected screens and fixed in place in front of the viewer. The total viewing range of a subject is called the field of view and is about 154 degrees horizontally and about 37 degrees vertically. Within this field of view, the eyes can register the objects surrounding the viewer (Fig. 1). The virtual terrain in our study consisted of a 5-km-long straight (or curved) stretch of road, grass, and trees with a mountain background. (10) Once the treadmill was attached to the PC system, the rate of the subject's movement matched the environmental flow on the screen at a rate of 30 frames per second. The VR programs also offered immediate biofeedback on the subject's condition with free-run (ie, continuously displayed without recording) HR, respiratory rate, and electromyography waveforms (I-330-C2 ([dagger])) of both thighs on an additional window at the right upper corner of the screen.

[FIGURE 1 OMITTED]

In some VR experiments, the users wear head-mounted displays and thus experience "total immersion" in VR. We chose to use large "wraparound" screens instead of head-mounted displays because head-mounted displays produce "simulation sickness" in some users. In addition, head-mounted displays prevent users from being able to see their own bodies, and this situation, for some, can be a highly unpleasant experience. (4)

Cardiorespiratory Testing

Cardiorespiratory testing was done initially to find each subject's HRmax and V[O.sub.2]peak. These data were used to determine the specific goals for the subject to reach during the subsequent exercise training sessions. We described cardiorespiratory testing in previous reports. (7,9) The anaerobic threshold, which is the point at which the steepness of the slope of the curve for V[O.sub.2] versus work output lessens as anaerobic glycolysis begins, also was measured.

In brief, for the exercise testing protocol used in the current investigation, subjects took part in up to 10 exercise periods of 3 minutes' duration. This protocol was identical for each subject. The grades and speeds of the periods are shown in Table 1. The subjects exercised in accordance with the protocol until they reached a level at which they experienced subjective exhaustion or a plateauing of their oxygen intake or until some clinical contraindication set in.

Training Protocol

All subjects were asked to perform submaximal endurance training exercises twice per week for about 3 months in the hospital until they were able to achieve a level that corresponded to 85% HRmax, 75% V[O.sub.2]peak, or both, as recorded during the exercise testing session. These target intensities were chosen on the basis of guidelines from the American College of Sports Medicine (11,12) and, being 15% to 25% lower than those used for subjects who are healthy, were thought to be safe for the compromised group in our study. Heart rate, V[O.sub.2], and treadmill grades and speeds were recorded for each subject during each session. Blood pressure was measured in these sessions every 5 minutes, before an increase in the treadmill grade or speed. The subject was connected to an electrocardiography machine for recording of HR. To estimate 75% V[O.sub.2]peak during the endurance training session, we recorded the speed and incline of the treadmill at 75% V[O.sub.2]peak during the exercise testing session and used attainment of this speed and incline as evidence that 75% V[O.sub.2]peak had been achieved. (13) This procedure has been used for years in exercise training, because it is impossible to use the precise procedure (use of a mask) to record oxygen and carbon dioxide levels. Oxygen consumption was chosen as a second exercise endpoint because many of the subjects were taking beta-adrenergic blockers, which affect HRmax. In the training protocol, the subject exercises on a treadmill, starting with 3 minutes of very low-level exercise (warm-up), after which the work rate is increased every 5 minutes, as shown in Table 1. In this way, the work rate adds 1% and 0.8 km/h (0.5 mph) during each 5-minute period until the subject reaches a score of 16 for perceived level of exertion on the Borg 6- to 20-point rating scale (that is, an effort somewhere between "hard" and "very hard") or until the subject reaches the target HR and V[O.sub.2].

Depending on the subject's condition, the training may be stopped before 30 minutes or may last a little longer than 30 minutes. In addition, the work rate increment is adjusted downward if required by the subject's cardiorespiratory status; that is, if, during a given work rate increment, a subject's blood pressure increased to more than 220 mm Hg systolic or 120 mm Hg diastolic or a subject's HR increased more than 10 beats per minute, (11) the therapist decreased the next slope or speed increment somewhat from the preset protocol. We did not always insist on the preset protocol for ethical reasons. Previous research also showed that changing the size or length of the increments in an exercise protocol does not change the maximum load that a subject is able to tolerate. (14)

Only the group 2 subjects were provided with the VR environment and asked to focus on a virtual scene. The group 1 subjects performed treadmill walking at measured speeds and grades without VR. In all other respects, the 2 programs were identical.

Data Analysis

Data entry and statistical analyses were performed with SPSS version 10.0. ([double dagger]) The 2-sample t test for independent samples was used to compare the baseline characteristics of the 2 groups. With the Kaplan-Meier method, the number of training sessions that the subjects in each training group underwent before the target event occurred was analyzed to evaluate the effect of VR on the time required to achieve the target goals. This method is used to analyze discontinuous "yes/no" events and takes into account the phenomenon that some subjects may exit a study before the target event has been reached. We defined the event to be evaluated as the achievement of target exercise capacity (ie, 85% HRmax, 75% V[O.sub.2]peak, or both). That is, the value of the response variable was equal to 1 if the target aerobic power was reached and equal to 0 if the subject failed to achieve the specified target. Comparisons of the resulting exercise goal achievement graphs (VR versus non-VR) were based on the log-rank and Breslow tests. The acceptable level for statistical significance was set at P<.05.

Results

Of the 24 subjects selected, 4 were found to be unable to tolerate the endurance training sessions because of their medical conditions, which were as follows: unstable angina (n= 1), uncontrolled symptomatic heart failure (n=2), and uncontrolled cardiac arrhythmias causing symptoms or hemodynamic compromise (n=l) (Fig. 2). These subjects therefore were excluded from the analyses, leaving 20 outpatients (20 men) in the study. The clinical characteristics of this final sample are shown in Tables 2 and 3. Five subjects (4 in the control group and 1 in the VR group) left the study before their second target goal had been achieved. One subject in the control group did not attain either target goal even after a maximum number (32) of training sessions.

[FIGURE 2 OMITTED]

There were no significant differences between the 2 subject groups in age, changes in body mass index from before surgery to after surgery, time lag between surgery and exercise testing, or time lag between surgery and intervention (Tab. 2). Subjects' clinical details regarding medications taken, the type of CABG, and the cardiac risk factors of smoking, hypertension, and diabetes mellitus are shown in Table 3.

Figures 3 and 4 show the percentage of subjects in each group who attained the target goals versus time during consecutive rehabilitation sessions under the VR and non-VR protocols. Cross marks on the graphs indicate sessions in which a subject dropped out without attaining the particular goal being measured on the graph and after which he was not included in the calculations for that goal. There was a significant difference between the VR and non-VR groups in the number of sessions required to achieve each target goal. To reach the target 85% HRmax, 9 of 10 subjects in the VR group needed no more than 9 training sessions, whereas by session 20, only 4 subjects in the non-VR group had achieved this target goal (Breslow test, P=-.0006; log-rank test, P=.0005) (Fig. 3). With regard to the target V[O.sub.2] (75% V[O.sub.2]peak), all 10 subjects in the VR group had achieved this target goal by the second round of training, but 15 training sessions were needed for a comparable number of subjects in the non-VR group to achieve this goal (Breslow test, P=.0013; log-rank test, P=-.0003) (Fig. 4).

[FIGURES 3-4 OMITTED]

The treadmill speeds and grades were recorded during the endurance training sessions to evaluate whether the VR experience affected the maximum work rate achieved in these sessions. Our findings showed that the subjects in the VR group achieved a significantly greater highest speed than the subjects in the non-VR group achieved ([bar.X] [+ or -] SD=4.64 [+ or -] 1.40 versus 3.70 [+ or -] 0.81 mph; P=.037) as they completed the submaximal endurance training exercise session.

Discussion

In this study, we investigated whether the incorporation of a VR experience into a cardiac rehabilitation exercise program would enable subjects to reach performance goals in fewer exercise sessions. The subjects in the VR group showed an increase in the maximum workload achieved during these sessions, a decrease in the number of sessions required to reach the target HR, and an extreme reduction in the number of sessions required to reach the target V[O.sub.2]. How might these results be explained?

Differences Between the Groups in the Slope and Speed Adjustments in the Training Protocol

Except for the incorporation of the VR experience, the exercise protocols were the same for both groups of subjects. However, it is worth noting that even if the incremental steps used had been different, the results would not have been affected, because previous experiments by others showed that variations in these steps do not change the exercise level at which a subject's maximum exercise tolerance occurs. (14,15)

Effect of the VR Experience Itself

The VR experience is a powerful one, quite different in intensity and quality from ordinary attention diversions, such as listening to music or watching television. Although participants know that the VR experience is not real, they describe the experience as feeling immersed in the VR scene, as if they are actually there, and they behave as if the virtual is real. (4) This immersion is physiological as well as psychological. For example, if VR shows the participant that his next step will cause him to fall off a cliff, not only will he refuse to step forward "off the cliff," but also his HR will go up as in fear, even though he knows at the same time that the cliff is not real. (4) The most likely explanation for our results is that it was this feeling of presence and involvement in an alternate experience that allowed the subjects in the VR group to tolerate higher maximum exercise levels and reach their individual exercise goals in a shorter time period. Whether more common distractions, such as listening to music, also would accelerate attainment of these goals we do not know, because our protocol was not designed to address this question.

Encouragement From Biofeedback

One difference between the subjects in the VR group and the control subjects was that the former subjects could see their HR, respiratory rates, and electromyography results in the right upper corner of the VR screen. However, they were not told what their target HR was. It is possible that this opportunity for biofeedback was a factor in their superior performance. The possible explanations are not known at this time but will be investigated in a future study.

Heart Rate Versus V[O.sub.2] Goals

Two target goals were used in this study: 85% HRmax and 75% V[O.sub.2]peak. Both study groups reached the V[O.sub.2] goal in fewer training sessions than were needed to reach the HR goal. We used standard nomograms in this study to calculate the degree of V[O.sub.2] considered to be the equivalent of the HR goal. (16) However, despite this strategy, the V[O.sub.2] goal seems to have been set at a lower level than the HR goal. Factors unrelated to HR, such as ambient temperature or physical fitness, can alter V[O.sub.2] and the caloric cost of exercise. (17) For this reason, HR and perceived exertion rating are the parameters most frequently used for assessing exercise intensity levels. However, because we were concerned about the effect of the beta-adrenergic blockers that some of the subjects were taking on the HR results, we wanted to use another independent, objective estimate of exercise intensity; therefore, we included V[O.sub.2] as a second goal.

Effect of VR on V[O.sub.2] Results

The subjects in the VR group reached their V[O.sub.2] goal by the end of the second training period, a time that seems too short for any substantial physiologic change to have occurred. They also reached greater highest treadmill speeds than the control subjects. Presumably part of the reason for these results was that the VR experience lessened the subjects' awareness of actual physical discomfort, and so they exercised to a higher level than they otherwise would have chosen to do.

Possible Long-Term Benefit

We do not yet know the long-term clinical benefit of early achievement of exercise goals. However, at the moment, early achievement of these goals would not affect insurance reimbursement, because currently the number of training sessions reimbursed is based on the patient's risk level, as determined by the American Association of Cardiovascular and Pulmonary Rehabilitation classification rules, as follows: low risk level, up to 6 supervised training sessions; middle risk level, up to 18 such sessions; and high risk level, more than 36 such sessions. (18,19)

Conclusion

We cannot explain exactly how VR influences exercise performance. However, our study showed a powerful effect of a VR environment on the progress of cardiac rehabilitation and the results suggest that incorporating VR into rehabilitation programs will accelerate the maximum recovery of a patient's cardiovascular fitness.

This article was received October 19, 2005, and was accepted May 10, 2006.

References

(1) Hesse S, Schmidt H, Werner C, Bardeleben A. Upper and lower extremity robotic devices for rehabilitation and for studying motor control. Curr Opin Neurol. 2003;16:705-710.

(2) Wang P, Kreutzer IA, Bjarnemo R, Davies RC. A Web-based cost-effective training tool with possible application to brain injury rehabilitation. Comput Methods Programs Biomed. 2004;74:235-243.

(3) Morganti F, Gaggioli A, Castelnuovo G, et al. The use of technology-supported mental imagery in neurological rehabilitation: a research protocol. Cyberpsychol Behav. 2003;6:421-427.

(4) Sanchez MV, Slater M. From presence to consciousness through virtual reality. Nat Rev Neurosci. 2005;6:332-339.

(5) Burdea G, Coiffet P. Virtual Reality Technology. New York, NY: John Wiley & Sons Inc; 1994:273.

(6) Annesi JJ, Mazas J. Effects of virtual reality-enhanced exercise equipment on adherence and exercise-induced feeling states. Percept Mot Skills. 1997;85:835-844.

(7) Chuang TY, Chen CH, Chang HA, et al. Virtual reality serves as a support technology in cardiopulmonary exercise testing. Presence. 2003;12:326-331.

(8) Hartung GH, Rangel R. Exercise training in post-myocardial infarction patients: comparison of results with high risk coronary and post-bypass patients. Arch Phys Med Rehabil. 1981;62:147-150.

(9) Chuang TY, Sung WH, Lin CY. Application of a virtual reality-enhanced exercise protocol in patients after coronary bypass. Arch Phys Med Rehabil. 2005;86:1929-1932.

(10) Kim NG, Choong KY, Jae JI. A new rehabilitation training system for postural balance control using virtual reality technology. IEEE Trans Rehabil Eng. 1999;7:482-485.

(11) American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription. 7th ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2006:133-173.

(12) Wasserman K, Hansen JE, Sue DY, et al. Principles of Exercise Testing and Interpretation: Including Pathophysiology and Clinical Applications. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 1999:115-142.

(13) Metier CP, Pollock ML, Graves JF. Exercise prescription for the coronary artery bypass surgery patient. J Cardiopulm Rehabil. 1986;6: 85-103.

(14) Redwood DR, Rosing DR, Goldstein RE, et al. Importance of the design of an exercise protocol in the evaluation of patients with angina pectoris. Circulation. 1971;43:618-628.

(15) Arstilla M. Pulse-conducted triangular exercise-ECG test. Acta Med Scand. 1972;529:103-109.

(16) American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription. 5th ed. Baltimore, Md: Williams & Wilkins; 1995:153-240.

(17) American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription. 6th ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2000:137-164.

(18) Paul-Labrador M, Vongavanich P, Merz NB. Risk stratification for exercise training in cardiac patients: do the proposed guidelines work? J Cardiopulm Rehabil. 1999;19:118-125.

(19) Roitman JL, Lafontaine T, Drimmer AM. A new model for risk stratification and delivery of cardiovascular rehabilitation services in the long-term clinical management of patients with coronary artery disease. J Cardiopulm Rehabil. 1998;18:113-123.

* Microsoft Corp, One Microsoft Way, Redmond, WA 98052-6399.

([dagger]) J&J Engineering, 22797 Holgan Ct NE, Poulsbo, WA 98370.

([double dagger]) SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

TY Chuang, MD, is Attending Physician and Associate Professor, Department of Physical Medicine and Rehabilitation, Taipei Veterans General Hospital, No. 201, Shih-Pai Rd, Sec. 2, Taipei, 11217 Taiwan, and School of Medicine, National Yang-Ming University, Taipei, Taiwan. Address all correspondence to Dr Chuang at: [e-mail].

WH Sung, PhD, is Assistant Professor, Department of Biomedical Engineering, I-Shou University, Kaohsiung, Taiwan.

HA Chang, PT, MS, is Physical Therapist and Research Associate, Department of Physical Medicine and Rehabilitation, Taipei Veterans General Hospital.

RY Wang, PT, PhD, is Professor and Acting Chairman, Institute and Faculty of Physical Therapy, National Yang-Ming University.

Dr Chuang and Dr Sung provided concept/idea/research design and facilities/equipment. Dr Chuang provided writing. Ms Chang provided data collection and analysis. Dr Chuang and Ms Chang provided subjects. Dr Wang provided research consultation (including review of manuscript before submission).

This study was approved by the Ethics Committee of the Institutional Board of Taipei Veterans General Hospital.

This study was supported by a grant from the National Science Council (NSC 92-2314-B-075-030).

Table 1.
Treadmill Grade and Speed Levels Used in Exercise
Testing and Training Protocols

Testing Protocol (a)      Training Protocol (b)

            Speed, km/h               Speed, km/h
Grade (%)   (mph)         Grade (%)   (mph)

1           1.6 (l) (c)   0           0.48 (0.3) (c)
0           1.6 (l)       0           0.96 (0.6)
0           2.4 (1.5)     1           1.76 (1.1)
3.5         3.2 (2)       2           2.56 (1.6)
7           4.8 (3)       3           3.36 (2.1)
5           4.8 (3)       4           4.16 (2.6)
7.5         4.8 (3)
10          4.8 (3)
12          4.8 (3)
15          4.8 (3)

(a) Each level was used for 3 minutes.

(b) Each level was used for 5 minutes.

(c) There was a 3-minute warm-up period.

Table 2.
Descriptive Data for Subjects Who Underwent Coronary Artery
Bypass Grafting (CABG) and Who Took Symptom-Limited Exercise
Tests With or Without Virtual Reality (VR)

                       Non-VR Group (a)   VR Group (a)
                       (n=10)             (n=10)

Parameter              [bar.X]    SD      [bar.X]    SD    P (b)

Age (y)                  63.70   10.03     65.70   14.48    NS

Increase in body          1.28    3.42      0.65    0.96    NS
  mass index (%) (c)

Months from date          3.21    3.18      2.81    3.38    NS
  of CABG to
  first evaluation

Months from date          4.19    4.37      4.47    3.47    NS
  of CABG to
  first intervention

V[O.sub.2]peak           19.50    3.75     18.06    3.47    NS
  (mL/kg/min) (d)

V[O.sub.2]               10.78    2.72      8.56    1.74    NS
  (mL/kg/min)
  at anaerobic
  threshold (d)

(a) All subjects in each group were men.

(b) NS=not significant; no significant differences were noted
between non-VR and VR groups, as determined by a 2-sample
independent t test (P > .05).

(c) Value recorded at the initial exercise test minus the value
recorded before surgery.

(d) Oxygen consumption ([VO.sub.2]) was determined from the
initial maximum exercise test. [VO.sub.2]peak=peak oxygen
consumption.

Table 3.
Clinical Details for Subjects (a)

                                No. of Subjects

                                Non-VR Group   VR Group
Parameter                       (n=10)         (n=10)

Smoking                         3              6

Diabetes mellitus               2              3

Hypertension                    4              6

Hyperlipidemia                  4              3

Medications
  ACE inhibitors                1              1
  Vasodilators                  2              3
  Diuretics                     2              1
  Beta-blockers                 4              4
  Calcium channel blockers      8              5
  Antithrombotic agents         9              5

Type of CABG
  Cardiopulmonary bypass pump   9              10
  Minimally invasive            1              0

(a) VR=virtual reality, ACE=angiotensin-converting enzyme,
CABG=coronary artery bypass grafting.