Effects of a developmental physical therapy program on oxygen saturation and heart rate in preterm infants.

Effects of a Developmental Physical Therapy Program on Oxygen Saturation and Heart Rate in Preterm Infants Physical therapy programs in neonatal intensive care units (NICUs) for preterm infants with chronic medical problems or abnormal neurological findings, or both, are designed to evaluate and treat developmental delays or abnormal patterns of movement. Developmental interventions in the NICU have included strategies to improve positioning, sensation, motor development, behavioral organization, feeding, and parent teaching. [1-3] All interventions must account for the fragile condition of preterm infants. The limited ability of many preterm infants to inhibit responses to excessive or poorly timed environmental stimuli can produce physiologic stress during routine medical and care-giving procedures. [4-6] Long et al found that hypoxia in infants with low birth weight was associated with various nursing procedures such as suctioning, diaper chanes, and injections. [4] Speidel reported a case study in which the arterial oxygen tension ([Pao.sub.2]) of a preterm infants dropped from a relatively stable average of 88 mm Hg to 58 mm Hg during routine vital sign monitoring. [5] Sweeney reported that preterm infants had significantly greater changes in heart rate in response to a standard neurological examination compared with full-term infants. [6] These findings support the practice of individualized intervention programs that carefully grade the amount and type of stimulation to the infant's level of tolerance.

A model proposed by Als et al considers the infant's responses to intervention and can assist the physical therapist with grading the intensity of treatment. [7] Adaptation and interaction with the environment are viewed as a continuous feedback process dependent on the infant's progression through a series of hierarchical stages of behavioral organization. The first stage is physiologic stability, followed by motor organization, modulation of behavioral state, and finally the ability to interact with the environment. This model proposes that the infant must regulate basic autonomic functions prior to developing motor competencies.

The preterm infants' tolerance for motor activity may be limited not only by their inability to organize behavioral responses, but also because of structural and functional immaturities of their cardiopulmonary system. [8] In comparison with the full-term infant, low lung compliance, higher pulmonary resistance, and relatively high chest wall compliance have been linked to excessie energy expenditure during resting tidal volume breathing in the preterm infant. [8] For these reasons, developmental physical therapy requires careful monitoring of physiologic stress signals to minimize the risk of placing excessive physical demands on preterm infants.

Percentage of arterial oxygen saturation ([Sao.sub.2]) and heart rate are two physiologic measures that can be monitored using a pulse oximeter. This device has gained recent attention and support as a reliable means of estimating [Sao.sub.2] in the adult and in the infant. [9,10] Some studies have demonstrated that when compared with transcutaneous oxygen monitors, pulse oximetry is both more accurate and clinically feasible. [11-13] Pulse oximetry has been used successfully to monitor exercise tolerance and the supplemented oxygen requirements of adults with cardiopulmonary disease. [14,15] Because developmental physical therapy intervention programs emphasize movement, increased activity, and presumably increased oxygen demands, pulse oximetry may be an objective method to monitor cardiorespiratory status and determine an infant's ability to tolerate physical therapy.

The purpose of this study was to document the [Sao.sub.2] and heart rate responses of a group of medically stable preterm infants to a developmental physical therapy intervention. Percentage of arterial oxygen saturation and heart rate measurements were taken before, during, and after six treatment activities performed in the side-lying and supported-sitting positions. The following research questions were addressed:

1. Are there significant changes from mean baseline levels of [Sao.sub.2] and heart rate as a result of the developmental physical therapy intervention?

2. Are the changes from mean baseline levels of [Sao.sub.2] and heart rate specific to either the side-lying or supported-sitting position?

3. Are the changes from mean baseline levels of [Sao.sub.2] and heart rate a function of the duration of the intervention?

4. What is the trend of the [Sao.sub.2] and heart rate measurements during the 15-minute period following the intervention?

Method

Subjects

Fourteen medically stable preterm infants born between 31 and 34 weeks gestational age were studied at 34 to 38 weeks postconceptional age (Tab. 1). Infants eligible for the study were stable on room air and did not have a cardiac defect, congential anomalies, or neurological dysfunction. Over a five-month period, all infants who met the eligibility criteria and for whom informed consent was received were included in the study. At the time of the study, the infants were not on medication at a dosage that would alter their behavioral state, had received no invasive medical procedures in the last two hours, and had a resting level of [Sao.sub.2] within a clinically safe range as determined by a neonatologist. Infants were recruited from the intermediate NICUs of Hahnemann University Hospital and Pennsylvania Hospital, both in Philadelphia, Pa.

Instrumentation

A Nellcor N-10 portable pulse oximeter (*1) was used to measure [Sao.sub.2] and heart rate. The N-10 pulse oximeter is a hand-held, battery-operated device that produces a print-out of [Sao.sub.2] and heart rate. The accuracy of the N-10 pulse oximeter was evaluated by Hess et al under conditions of normoxia and hypoxemia. [16] Correlation of N-10 pulse oximeter measurements with arterial blood samples analyzed on a co-oximeter was .96. The standard error of measurement was 2% in the [Sao.sub.2] range of 77% of 100%.

Procedure

The repeated-measures research design consisted of three phases--baseline, intervention, and recovery--during which [Sao.sub.2 and heart rate were measured. To account for differences in preterm infant motor development, a criterion-referenced assessment was designed to determine each infant's abilities to engage them in an appropriate treatment plan. The assessment contained six items that measured spontaneous face clearing in the prone position, antigravity movements of the extremities, and head control. One of the two standardized treatment plans (Appendix) was selected based on the infant's score. Interventions were performed in the side-lying and supported-sitting positions, but differed in intensity and level of difficutly. All except two of the infants participated in the more advanced activities (Treatment B).

To prepare for data collection, a disposable oxisensor was applied to the lateral border of the infant's foot, and then the leads were connected to the preamplifier of the pulse oximeter. The accuracy of the oxisensor was verified by checking the reading of the oximeter with that of a cardiorespiratory monitor, also attached to the infant. Throughout all three phases of the study, a measurement was considered valid if the difference in heart rate between the pulse oximeter and cardiorespiratory monitor differed by no more than 10 beats per minute (bpm). To further ensure that the measurements were free of movement artifact, the infant's leg with the oxisensor was stabilized using light pressure while each recording was taken. The first author (MKK) performed the intervention and was blind to all pulse oximeter readings. An assistant operated the pulse oximeter, and had two consecutive Sao.sub.2 readings been less than 80% the investigator would have been informed to stop the intervention. A minimal level of 80% was selected based on the recommendation of the neonatologists at the institutions.

The eight-minute baseline phase began with the infant positioned prone while in a "quiet alert" or "quiet awake" state, as defined by Dubowitz and Dubowitz. [17] If the infant began to cry, the procedure was stopped. Measurements were recorded every minute.

During the intervention phase, six treatment activities were administered to each infant (Appendix). Three activities were performed in the side-lying position and three in supported-sitting position. Each activity was performed for one minute. The activities chosen were representative of developmental physical therapy procedures. The side-lying position was selected because it eliminates gravity and encourages relaxation, visual awareness, and midline orientation of the upper extremities. the supported-sitting position places postural demands on the infant to control the head and trunk. The beginning position and sequence of activities within each position were randomly ordered. Measurements were taken at the conclusion of each treatment activity.

Because pulse oximetry is sensitive to motion artifact, the normal search time of 10 to 30 seconds was frequently increased during the intervention phase. To control for this variability, if a reading was not obtained after 60 seconds, the intervention resumed and the data point for the activity was considered missing. The number of missing data points represented 17% of the measurements.

Following the intervention phase, each infant was repositioned prone to begin the recovery phase. The pulse oximeter was put in the continuous mode for a 15-minute period, and measurements were taken every 30 seconds.

Data Analysis

To address the effect of intervention and intervention position on Sao.sub.2 and heart rate, each subject's mean Sao.sub.2 and heart rate were calculated for the baseline phase in the prone position (eight measurements) and for the intervention phase in the side-lying (three measurements) and supported-sitting (three measurements) positions. The grand means were then determined and analyzed for each position by a one-way analysis of variance (ANOVA) for repeated measures.

To address the effect of duration of intervention on Sao.sub.2 and heart rate, seven means for both Sao.sub.2 and heart rate were calculated for each subject: baseline (mean of eight measurements), early treatment (mean of first and second intervention phase measurements), middle treatment (mean of third and fourth intervention measurements), late treatment (mean of fifth and sixth intervention measurements), early recovery (mean of first five minutes), middle recovery (mean of middle five minutes), and late recovery (mean of last five minutes). The grand means for each variable were analyzed by a one-way ANOVA for repeated measures. For all ANOVAs, post boc analyses of significant effects were performed using the Newman-Keuls method of multiple comparisons. The .05 level of significance was used for all statistical tests. Descriptive statistics were used to address the trend of the Sao.sub.2 and heart rate measurements recorded at 30-second intervals during the 15-minute recovery period.

Results

The mean values of Sao.sub.2 and heart rate for the baseline phase and for the side-lying and supported-sitting positions of the intervention phase are listed in Tables 2 and 3, respectively. No significant differences in mean Sao.sub.2 were found for the baseline, side-lying, and supported-sitting measurements (F = 1.96; df = 2,26; p = .16). A significant effect was found for heart rate (F = 12.72; df = 2,26; p [is less than] .0001) with a mean increase from the baseline phase of 10 bpm in the side-lying position and 11 bpm in the supported-sitting position of the intervention phase. Post hoc comparisons indicated that the mean heart rates for both side-lying and supported-sitting positions in the intervention phase were significantly higher than the mean heart rate for the baseline phase (p [is less than] .05). The mean heart rates for the side-lying and supported-sitting positions of the intervention phase were not significantly different.

The mean values of Sao.sub.2 and heart rate for the seven time intervals of the baseline, intervention, and recovery phases are listed in Tables 4 and 5, respectively. There were no significant differences in mean Sao.sub.2 among the seven time intervals (F = 1.64; df = 6,66; p = .15). A significant effect was found for heart rate (F = 6.72; df = 6,66; p [is less than] .0001). Post boc comparisons indicated the mean heart rates for the early, middle, and late treatment intervals were all significantly higher than mean baseline heart rate and the mean heart rates for the three recovery phase time intervals (p [is less than] .01). Neither the mean heart rates for the three treatment phase intervals nor the mean heart rates for the baseline and recovery phases differed significantly.

The trends of the mean Sao.sub.2 and heart rate values during the baseline and recovery phases are shown in Figures 1 and 2. The means for Sao.sub.2 were higher for the baseline phase than for the recovery phase; however, none of the means for the recovery phase were less than 90%. There was a considerable overlap in mean heart rate during the baseline and recovery phases.

Discussion

The preterm infants in this study responded to the developmental physical therapy intervention with a significant increase in heart rate, but no significant change in Sao.sub.2. The mean heart rate increased 10 and 11 bpm during intervention in the side-lying and supported-sitting positions, respectively, representing an increase of approximately 6% to 7% over the mean baseline value. Heart rate quickly returned to baseline values during the recovery phase. In the adult, a normal response to submaximal exercise is a linear increase in heart rate and cardiac output with increases in work load and oxygen consumption. [18] Although this tenet has not been tested in premature infants, heart rate has been shown to correlate closely with energy expenditure in newborn infants. n19 Within this context, the increased mean heart rate during the intervention phase suggests that the activities were of sufficient intensity to increase energy expenditure, hence to produce an "exercise effect." The infants' cardiovascular response, coupled with their ability to maintain arterial oxygenation during the intervention phase, suggests that they tolerated the treatment activities without compromising their physiologic stability.

Within limits, an increase in heart rate may result in augmented cardiac output, increased pulmonary blood flow, recruitment and distention of pulmonary capillaries, decreased pulmonary transit time, and increased pulmonary capillary surface area for gas exchange. [20] In conjunction with a host of other potential mechanisms, the increase in heart rate noted in this study may contribute to the maintenance of arterial oxygenation and support oxygen delivery to meet the presumably "exercise-included" increased oxygen demands. Beyond a physiologic limit, increases in heart rate result in a decreased cardiac output secondary to ineffective ventricular filling as diastolic filling time becomes shortened. [21] If myocardial contractile reserve is limited, stroke volume, cardiac output, and pulmonary blood flow may subsequently decrease. [21] This decrease would suggest that careful monitoring of [Sao.sub.2] and heart rate during exercise is necessary to guard against potential cardiopulmonary compromise and oxygen desaturation.

Clinical Implications

The model of behavioral organization proposed by Als et [al.sup.7] is applicable to physical therapy because it considers the need for developmental progression based on physiologic stability. The regulation of [Sao.sub.2] and heart rate during physical activity may serve as an indication that the infant can make the transition to the next developmental stage in the hierarchy without excessive compromise of other systems. The infants in this study demonstrated appropriate physiological responses and apparently were able to adapt to the demands placed on them during the intervention. This response would indicate that the infants could sufficiently regulate autonomic functions and begin to develop motor competencies.

For the infants in this study, treatment progression would initially involve similar activities, while increasing the demand of active participation by the infant. As the intensity of intervention is graded to accommodate the infant's behavioral responses, the cardiorespiratory responses must be monitored as well. As motor organization improves, the differentiation of responses allows the infant to interact in a more effective and safe manner.

The time interval between treatment activities may have affected the results and is an important consideration for treatment planning. There was a 12-to 60-second pause between each treatment activity to allow accurate recording with the pulse oximeter, providing the infants with a rest period between activities. The interval between the end of each treatment activity and the pulse oximeter readings precluded measurement of transient changes in [Sao.sub.2] and heart rate immediately after each treatment activity. Regardless of the time-lag, however, all infants presumably demonstrated an exercise effect, as reflected by the increased heart rate. Lastly, the interval between each activity varied among the infants (12-60 seconds). Despite this potential influence on the results, the rest periods may have been fortuitous and may have contributed to an intervention that was appropriate in intensity and duration. Individual programs that are guided by cardiopulmonary responses may promote, rather than compromise, physiologic stability.

The results for [Sao.sub.2] and heart rate during the 15-minute recovery period indicated that the infants remained relatively stable following intervention. A duration of 15 minutes was chosen specifically to address the physiologic changes that may occur when physical therapy is completed and the infant is left to "rest." It is noteworthy that one infant demonstrated several low [Sao.sub.2] readings during the recovery period. This finding is important, particularly in infants who are treated for a longer period of time, because clinical signs of hypoxia are masked until low levels of [Pao.sub.2] are reached. Interventions, whether care-giving or therapeutic, must be planned and monitored so that the infant is given a sufficient number of rest periods and time for recovery prior to subsequent interventions.

As an adjunct to conventional cardiorespiratory monitors, pulse oximetry provides an objective measurement of the infant's tolerance to stimuli and therefore assists the physical therapist in selecting treatment activities that are appropriate from both a developmental and a physiological perspective. The major drawback to pulse oximetry, especially in a study of this nature, is its sensitivity to movement. Because pulse oximetry relies on pulsatile fluctuations in the intensity of transmitted light, measurement is adversely affected by movement. Although the Nellcor pulse oximeter has been reported to be less sensitive to movement artifact compared with two other oximeters (KJ Barrington, NN Finer, A Ryan; unpublished data; 1986) and despite attempts to minimize movement artifact, there were numerous missing data points in this study. The simultaneous comparison of the heart rate recording from the pulse oximeter with the heart rate recording from a cardiorespiratory monitor, as recommended by Barrington and colleagues, enhanced the validity of the results obtained in this study.

Conclusion

Although researchers have demonstrated adverse physiologic responses to routine care-giving activities and medical procedures, little has been done to objectively document the effects of developmental physical therapy activities in the preterm neonate. The benefits and risks of physical therapy in terms of developmental gains and induced physiologic stress are critical issues for justifying intervention with preterm infants. The results of this study indicate that healthy preterm infants responded to six developmental activities with a significant increase in mean heart rate, but no significant change in mean [Sao.sub.2]. The increase in heart rate was not affected by treatment position or duration. During a 15-minute recovery phase, mean heart rate returned to baseline values, and there was little change in mean [Sao.sub.2]. With appropriate attention to minimizing movement artifact, we have found pulse oximetry to be a feasible method of monitoring physiologic status and recommend its use during and following intervention with preterm infants in the neonatal nursery. Future studies are recommended to examine when preterm infants are able to safely tolerate sensorimotor activity and to establish guidelines for the intensity and duration of developmental interventions. A better understanding of the physiologic consequences of intervention, particularly for the high-risk preterm infant, will enable physical therapists to more objectively weigh the potential therapeutic benefits of treatment against the physiologic risks.

Acknowledgments

We would like to acknowledge committee members Mary Watkins, MS, Michael Magno, PhD, and Johanna Deitz, MS, for their advice and assistance; David Metzger, PhD, Mary Lazar, MS, and Harvey Kushner, PhD, for their statistical assistance; Marjorie Curtis-Cohen, MD, and the Hahnemann Hospital nursery staff; Vinod K Bhutani, MD, and the Pennsylvania Hospital nursery staff; and Jeffrey Kelly for his assistance with data collection.

(*1) Nellcor Corp, 5495 Whitesell St, Hayward, CA 94545.

References

[1] Dickson JM: A model for the physical therapist in the intensive care nursey. Phys Ther 61:45-48, 1981

[2] Anderson J, Auster-Liebhaber J: Developmental therapy in the neonatal intensive care unit. Physical and Occupational Therapy in Pediatrics 4(1):89-106, 1984

[3] Sweeney JK: Neonates at developmental risk. In Umphred DA (ed): Neurological Rehabilitation. St Louis, MO, C V Mosby Co, 1985, vol 3, pp 137-164

[4] Long JG, Philip A, Lucey J: Excessive handling as a cause of hypoxemia. Pediatrics 65:203-207, 1980

[5] Speidel B: Adverse effects of routine procedures on preterm infants. Lancet 1:864-866, 1978

[6] Sweeney JK: Physiologic adaptation of neonates to neurological assessment. In Sweeney JK (ed): The High Risk Newborn: Developmental Therapy Perspectives. Binghampton, NY, Hawthorne Press, 1986, pp 155-169

[7] Als H, Lester B, Brazelton TB: Dynamics of behavioral organization of the premature infant: A theoretical perspective. In Field TM, et al (eds): Infants Born at Risk: Behavior and Development. New York, NY, Spectrum Publications Medical and Scientific Books, 1979, pp 173-192

[8] Guslits BG, Gaston GE, Bryan MH, et al: Diaphragmatic work of breathing in premature human infants. J Appl Physiol: Respirat Environ Exercise Physiol 62:1410-1415, 1987

[9] Yelderman M, New W: Evaluation of pulse oximetry. Anesthesiology 59-349-352, 1983

[10] Jennis MS, Peabody JL: Pulse oximetry: An alternative method for the assessment of oxygenation in newborn infants. Pediatrics 79:524-528, 1987

[11] Deckhardt R, Steward D: Non-invasive arterial hemoglobin oxygen saturation versus transcutaneous oxygen tension monitoring in the preterm infant. Crit Care Med 12:935-939, 1984

[12] Ramanathan R, Durand M, Larrazabal C: Pulse oximetry in very low birthweight infants with acute and chronic lung disease. Pediatrics 79:612-617, 1987

[13] Solimano AJ, Smyth JA, Mann TK, et al: Pulse oximetry advantages in infants with bronchopulmonary dysplasia. Pediatrics 78:844-849, 1986

[14] Smyth RJ, D'Urzo AD, Slutsky AS, et al: Ear oximetry during combined hypoxia and exercise. J Appl Physiol: Respirat Environ Exercise Physiol 60:716-719, 1986

[15] Smoker JM, Hess Dr, Frey-Zeiler VL, et al: A protocol to assess oxygen therapy. Respiratory Care 31-35-39, 1986

[16] Hess D, Kochansky M, Hassett L, et al: An evaluation of the Nellcor N-10 portable pulse oximeter. Respiratory Care 31:796-802, 1986

[17] Dubowitz L, Dubowitz V: The Neurological Assessment of the Preterm and Fullterm Infant. Philadelphia, PA, J B Lippincott Co, 1981

[18] Irwin S: Abnormal exercise physiology. In Irwin S, Tecklin JS (eds): Cardiopulmonary Physical Therapy. St Louis, MO, C V Mosby Co, 1985, vol 1, pp 50-63

[19] Chessex P, Reichman BL, Verellen GJ, et al: Relation between heart rate and energy expenditure in the newborn. Pediatr Res 15:1077-1082, 1981

[20] Guyton AC: Textbook of Medical Physiology, ed 7. Philadelphia, PA, W B Saunders Co, 1986

[21] Vidyasagar D, Sarnaik AP (eds): Neonatal and Pediatric Intensive Care. Littleton, MA, PSG Publishing Co Inc, 1985

MK Kelly, MS, PT, is Clinical Research Specialist, Children's Seashore House at Children's Hospital of Philadelphia, 34th & Civic Center Blvd, Philadelphia, PA 19104 (USA).

R Palisano, ScD, is Assistant Professor, Program in Physical Therapy, Hahnemann University, Philadelphia, PA 19102-1192.

M Wolfson, PhD, is Assistant Professor of Physiology and Pediatrics, Temple University School of Medicine, Philadelphia, PA 19140.

This study was completed in partial fulfillment of the requirements for Ms Kelly's Master of Science degree in the Pediatric Physical Therapy Program at Hahnemann University.

This study was partially supported by Grant #G008530031 to Ms Kelly from the US Department of Education and by grants to Dr Wolfson from the Foundation for Physical Therapy and the American Heart Association.

This article was submitted October 18, 1988; was with the authors for revision for fire weeks; and was accepted February 1, 1989.