Infants born prematurely frequently have motor delays through the first 12 to 24 months of life.1–4 Delays may be present in development of the ability to sit independently and rotate the trunk.3,4 Infants born prematurely also may have increased co-contraction of hip, leg and trunk muscles for increased postural stability,3 decreased quality of movement,1,3 and decreased reaching abilities2 compared to infants born at term. Once they reach school age, children who were born prematurely are more likely than their peers to be clumsy and to have poor visual motor5 and fine motor6 skills. Developmental delays and qualitative motor impairments seen in children who were born prematurely may be related to abnormalities in active muscle power and postures during the first months of life.4
Little quantitative research has been conducted on variations in resting trunk position between infants born preterm and infants born at term. Dargassies,7 Casaer,8 and Gesell9 described the postures of infants born at term suggesting that healthy term infants maintain their trunks in flexion while they are awake and active. These early researchers,7–9 however, used descriptive methods to characterize resting postures and did not provide sufficient information about their methods to permit replication of their work. Much of the quantitative research about infants’ trunk and extremity muscle tone as measured by passive resistance to lengthening the muscle10 is based on findings from standard neurological or newborn exams which focus on an infant’s movements in response to handling.7,10 Based on these exams, infants born preterm are thought to have less flexor muscle tone and to spend more time in an extended position than infants born at term. The amount of flexor tone is reported to increase as gestational age increases.7,10 Comparisons between infants born preterm and at term indicate that infants born preterm exhibit less flexion during elicited movements and in resting positions than infants born at term.7,10 Neurological and newborn assessments used in these studies, however, do not provide specific information about positioning of the infant’s trunk during active movements.
Recently, researchers11,12 investigating variation in postures have reported discrepancies between their results and those of Dargassies7 and Casaer.8 These discrepant results may be at least partially explained by differences in the behavioral states of the infants during the assessments.11 Dargassies7 assessed infants while they were awake and active. Cioni et al12 and Vles et al11 studied infants in Prechtl state 1 or 212 or in a wide range of states.11 Sleep affects the muscle tone and posture of infants born at term, and is likely to have an effect in infants born preterm as well.8 Given the variations in posture and muscle tone associated with different behavioral states, researchers must take behavioral state into account when examining infant postures.
Trunk flexion during active movement in a supine position has not been assessed in a quantitative or reproducible manner in infants born either at term or preterm. Although differences have been reported in the muscle tone of infants born preterm and infants born at term on neurological examinations, extremity position or muscle tone observed during manipulation of the infant’s body may not provide a valid representation of posture during spontaneous active movements. In addition, the possible progression from excessive trunk extension to postural instability and childhood motor, visual perceptual, cognitive, and psychological delays has not been investigated. Quantitative documentation of increased trunk extension and decreased trunk flexion in infants born preterm is needed as a basis for further studies of the relationship between trunk postures in young infants and developmental delays during the first few years of life.
The primary purpose of this study was to compare trunk position in supine of infants born preterm and at term. A secondary purpose was to determine the feasibility of using pressure data to assess trunk position in this population.
Four research questions relating to the primary purpose of the study were addressed: For infants positioned in supine at 38 to 43 weeks gestational age and in an alert (state 4) or active alert (state 5) behavioral state,13 1) Do infants born at term maintain their trunks in flexion or neutral (non flexed or extended position) more than two-thirds of the time? 2) Do infants born preterm maintain their trunks in flexion for a smaller percentage of the time than infants born at term? 3) Do infants born preterm maintain their trunks in extension for a larger percentage of the time than infants born at term? and 4) Do infants born preterm maintain their trunks in flexion for shorter durations per flexion event than infants born at term?
Based on previous literature7–9 and the results of unpublished pilot studies conducted by the authors, infants born at term were expected to spend less time in extension as a result of strong flexion tendencies. Thus, they were expected to spend less than one-third of the time in extension, resulting in more than two-thirds (>67%) of the time being in flexion or neutral. Infants born preterm were expected to spend less total time in flexion and, once in a flexed position, to maintain that position for shorter time periods than infants born at term.
Thirty-three infants were included in this study. Eighteen infants were born preterm (less than or equal to 37 weeks of gestation) and fifteen were born at term (38 to 42 weeks of gestation) (Table 1). All subjects were born at weights that were appropriate for their gestational age (AGA). Infants born preterm were assessed at 41 to 43 weeks gestational age, while infants born at term were assessed one to three weeks after delivery.
Subjects were recruited from Wake Medical Center, University of North Carolina Hospitals and the surrounding communities. Infants were identified by staff at the birth hospital through fliers posted at the hospitals, and by personal contacts of the investigators and study participants. Inclusion criteria determined from parental report and/or the infant’s medical record were: 1) no diagnosed neurological abnormalities, 2) no periventricular leukomalacia (PVL) or grade 3 or 4 interventricular hemorrhage (IVH), if ultrasound or other brain imaging study was completed, 3) no seizures, 4) no congenital abnormalities, 5) no diagnosed genetic or endocrine syndromes, 6) no need for supplemental oxygen at the time of assessment, 7) no drug or alcohol exposure, and 8) parent(s) able to speak English.
Parents who were interested in having their infant participate were contacted by telephone or in person by the principal investigator (PI). A screening questionnaire was completed with the family. The parent’s responses to the questions on the screening questionnaire were confirmed during a review of the infant’s medical record when possible. If the infant was eligible to participate, a data collection session was scheduled at the infant’s home. All data collection was completed by the PI, who is certified in administration of the Neonatal Behavioral Assessment Scale (NBAS).13 Parents of infants who participated in the study received a ten-dollar gift certificate at the completion of the assessment.
Of the 18 infants in the study who were born preterm, one infant had a grade 1 IVH, five infants were on a ventilator for at least one day (range one to 60 days, mean 18.18 days), and one infant had been diagnosed with chronic lung disease, but did not need supplemental oxygen at the time of the assessment. All term infants were on room air and their parents reported they had been healthy since delivery.
A firm closed cell foam mat was placed on the floor. An Ultra Thin Seat Mat* was placed on top of the foam mat and covered by a thin blanket. The Seat Mat was 0.36 mm thick and measured pressure at 16 × 16 points in a 43 cm2 sensing area. The Ultra Thin Seat Mat was connected via an interface module to a laptop computer. The Seat Mat software (FSA version 3.1.027) was used to acquire pressure data at a rate of 5 Hz. Test-retest reliability of the Ultra Thin Seat Mat was assessed by placing a known weight on a 3 cm square area of the pressure mat. The pressure exerted by each weight (range 3.5 to 10.5 ounces) was recorded for 6000 frames (20 minutes) on two separate occasions on two consecutive days. Interclass correlation coefficient (ICC) model 3,1 was used to compare the pressure data obtained for the two occasions. The ICC value was 0.99.
The PI met with the infant and at least one parent or legal guardian. To facilitate collection of data for both sleep and wake states, the test session was scheduled at a time when, according to the caregiver, the infant was likely to be transitioning from a sleep to an awake state or vice versa. The procedures of the study were reviewed with the parent. The parent signed an informed consent form approved by the Committee on the Protection of the Rights of Human Subjects at the University of North Carolina at Chapel Hill or Wake Medical Center, depending on the recruitment location. Assessment details including time, date, and infant’s starting position on the pressure mat were recorded on the record sheet along with time since last feeding and last nap.
The parent placed the infant supine on the pressure mat. The infant wore a diaper and a thin sleeper or t-shirt. The locations of the infant’s head and pelvis were recorded in reference to the pressure mat’s grid system. Pressure data were collected during two segments, one segment when the infant was in behavioral state 1 or 2 (sleep segment) and one segment when he/she was in behavioral states 4 or 5 (awake segment).13 The purpose of collecting the sleep segment data was to determine the infant’s baseline or neutral trunk position which would be compared with the data from the awake segment. An infant’s neutral trunk position was the relaxed position of his/her trunk while asleep, representing the mid-range of trunk flexion and extension. The data included in the sleep segment were collected for thirty seconds at any time the infant fell asleep during the data collection session. The infant’s sleep segment could be collected at the beginning, in the middle of, or after the awake segment, depending on the infant’s sleep-wake cycle. When necessary, parents facilitated the infant to fall asleep by feeding, holding, or providing a pacifier. We were unable to obtain sleep data for one infant born preterm; therefore, this infant’s data were not included in the analysis.
For the awake segment, collection of data from the pressure mat was initiated at a time when the infant was on the mat and awake, in behavioral state 4 or 5. The infant’s parent stood near the infant to insure that the infant did not roll off the mat and to encourage the infant to stay awake. The parent was permitted to talk to or look at the infant and present toys or pacifiers as needed to keep the infant happy and awake. When presenting toys to the infant, the toys were held in a stationary position within the infant’s field of view. The infant was not encouraged to turn his head or track the toy during the assessment. If the infant began to cry, the parent picked up the infant and calmed him/her. Once the infant was calm, the parent returned the infant to the pressure mat. Pressure data were recorded for a minimum of five minutes or 1500 total frames of alert state 4 or 5. The 1500 frames of the awake segment were collected in several shorter segments if the infant had difficulty remaining awake in supine without swaddling. One term infant was unable to remain in state 4 or 5 long enough to obtain 1500 frames of data. Therefore 1297 frames were used for analysis of the awake segment for this infant.
Throughout the recording, notations were made in the data collection software at each state change (Brazelton 1–5)13 and when the infant rolled out of a supine position, was picked up, or was repositioned on the pressure mat. All data collection sessions lasted less than 90 minutes, with most sessions requiring less than 45 minutes, including the setup of equipment in the family’s home.
This method of using pressure measurements to assess trunk position in infancy was developed for this study. Test-retest reliability of the measurement technique was examined for a sub-sample composed of the first five infants born at term whose parents agreed to repeated assessments. These five infants were retested within 36 hours of their first assessment, using procedures identical to those used in the first assessment.
Each infant’s pressure recording was reviewed and the sleep and awake segments were clearly identified based on notations entered in the data collection software during the assessment. Periods when the infant rolled out of supine were identified and excluded from the awake and asleep segments. Each data output screen depicted the pressure on each piezo receptor numerically and chromatically (Figure 1). One hundred and fifty frames of data collected during states 1 or 2 (sleep segment data) and 1500 frames of data collected during states 4 or 5 (awake segment data) were marked and exported to a spreadsheet for analysis. Frequently, the 150 sleeping and 1500 awake frames were from non-consecutive time periods. In these cases, the longest consecutive time periods were added together to achieve the required 150 or 1500 frames for the sleep and awake segments, respectively.
The pressure sensors representing the location of the head and pelvis were identified from notations made during data collection. Data from the sensors in the area between the head and the pelvis were considered to represent the trunk. The maximum pressure value for the array of sensors representing the locations of the head, trunk, and pelvis was determined for each frame. The ratio of head and pelvis pressure to trunk pressure was determined for each frame using the formula:
Equation (Uncited)Image Tools
The mean pressure ratio and standard deviation were calculated for the 150-frame sleep segment of each infant. Based on Casaer’s8 findings that sleeping infants relax their trunk, face, and extremities onto the support surface, the mean pressure ratio from the sleep segment for each infant was considered to represent the infant’s neutral trunk position. This value was used to categorize each frame of the awake segment as representing a neutral, flexed, or extended trunk position. Any frame in the awake segment that had a pressure ratio within two standard deviations of the mean pressure ratio for the sleep segment was categorized as representing a neutral trunk position. Frames from the awake segment that had pressure ratios greater than two standard deviations above the mean pressure ratio for the sleep segment were categorized as representing extension, while those with pressure ratios more than two standard deviations below the mean pressure ratio for the sleep segment were categorized as representing flexion. We used the relatively conservative criterion of a change of more than two standard deviations above or below the mean to insure that movements into flexion or extension were outside the range of normal variability in neutral in order to be considered as representing a position of trunk flexion or extension.
After each frame of the awake segment had been categorized as representing a neutral, flexed or extended trunk position, the total duration of trunk flexion and extension and the trunk flexion event duration were determined for each infant. Total duration was the total number of frames the infant’s trunk was in each position, expressed as a percentage of the total number of frames in the awake segment. Flexion event duration was determined by multiplying the number of consecutive frames that the infant’s trunk was in flexion (flexion burst) by the sampling period (200 msec).
Data from the five infants included in the reliability sub-sample were reduced using the same methods, except that both awake segments were compared to the same sleep segment ratio. Infants in the reliability sub-sample tended to have very restless sleep segments or short sleep segment data for one of their two data collection sessions. Data from the sleep segment with the lower standard deviation were used to categorize trunk positions for these five infants for both sessions.
ICCs (3,1) were calculated to compare the mean pressure ratios identified as representing flexion, extension and neutral for the two testing sessions for the five infants in the reliability sub-sample. These ICC values represent the reliability of the methods to categorize similar pressure ratios as the same position.
Test-retest reliability was assessed by calculating the total duration of flexion and extension and flexion burst duration on the two test occasions for each of the five infants included in the reliability sub-sample. The ICCs (3,1) for test-retest reliability include variations due to measurement error as well as true subject variability between sessions.
To address the first research question, the total duration of trunk flexion or neutral positioning in infants born at term was calculated as a percent of the total time. A one-sample t test for the mean was conducted to determine if the mean amount of time in flexion or neutral was greater than 67% (two-thirds) for the infants born at term. To address the second research question, the Mann-Whitney U test was conducted to compare the total duration of flexion between infants born preterm and infants born at term. To address the third research question, the total durations of extension in infants born preterm versus infants born at term were compared using the Mann-Whitney U test. To address the fourth research question, the Mann-Whitney U test was used to compare the mean duration of flexion events between infants born preterm and infants born at term. The significance level for all tests was set at 0.05. The limited number of tests performed and the complete reporting of the values of the test statistic and p-values in this study limits the need for correction or modification of the alpha level.14
Data were successfully collected on all but one infant included in this study. While assessing each infant, the position of the infant’s head and pelvis could be readily identified and recorded. Each frame in the awake segment could be clearly categorized as flexion, extension, or neutral.
The ICC values for the mean pressure ratios representing flexion, extension and neutral for the five infants born at term included in the reliability sub-sample were 0.84, 0.60, and 0.85 respectively. The ICC values for test-retest reliability of the total duration of flexion and, extension, and the flexion burst duration were 0.34, 0.40, and 0.30 respectively.
The mean sleep segment pressure ratio for infants born at term and those born preterm were 4.5 ± 1.57 and 3.41 ± 1.19, respectively. The mean pressure ratio values for all frames in the awake segment categorized as representing flexion, extension and neutral are presented in Table 2. Mean durations of either flexion or neutral trunk positioning for each group, expressed as a percent of the total awake segment, are illustrated in Figure 2. Infants born at term, but not preterm, spent significantly more than two-thirds of the awake segment in flexion or neutral (t = 2.479, df = 14, p = 0.027 for term; t = 1.504, df = 16, p = 1.52 for preterm, Figure 2). Mean duration of flexed and extended trunk positions for each group, expressed as a percent of the total awake segment, are illustrated in Figure 3. The preterm and term groups did not differ significantly in the total duration of either flexion or extension (U statistic = 157.50, p = 0.257 for flexion; and U Statistic = 97.50, p = 0.257 for extension; Figure 3, Table 3). The means and standard deviations of flexion event duration for the two groups are represented in Figure 4. Differences between infants born preterm and at term in the mean flexion event duration were not significant (U statistic = 149.0, p = 0.417; Figure 4, Table 3).
The infants born preterm included in this study were born at a wide range of gestational ages. The mean and standard deviations of each variable for infants born at less than 30 weeks of gestation and those born between 30 and 34 week of gestation are presented in Table 4. Because of the small number of infants born at less than 30 weeks of gestation, statistical comparisons were not performed.
The results of this study provide limited support for the common clinical observation that infants born preterm spend less time with their trunks in a flexed position when in supine than infants born at term. Although infants born at term, but not preterm, spent more than two-thirds of the time with their trunk in either a flexed or neutral position, no significant differences between the two groups in total duration of flexion or extension or in flexion event duration were observed. However, these results must be interpreted with care, as the variability seen within groups was large, especially for the term group.
The reliability data collected in this study on infants born at term indicate the methods used for categorizing trunk position are fairly reliable with ICC values ranging from 0.60 to 0.85. A single infant in the reliability sub-sample had much greater values for extension on the second day of testing when he was very fussy. This contributed to the slightly lower than anticipated ICC values for extension. Much lower ICC values for test-retest reliability provide evidence that the percent of time infants spend with their trunks in particular positions varies from day to day.
Subject variability may be related to the infant’s behavior during each assessment. Awake data were collected while the infants were in a state 4 or 5. However, these states encompass a broad range of activity levels, including alert with minimal motor activity, considerable motor activity, and fussy vocalizations.13 For example, an infant could be assessed one day while awake and happy with minimal motor activity and the following day be fussy and have a great deal of motor activity during the assessment. This could result in very different trunk positions and movement patterns between assessment sessions. The methodology could be improved by finding better ways to measure, record and control for small state changes that may affect trunk position. Furthermore, five minutes may be too short a time period to represent an infant’s characteristic trunk position. A longer data collection session might provide a better representation for each infant and result in decreased between- subject variability. In the future, it may be beneficial to collect a longer period of awake data on each infant. The reliability of this new methodology should continue to be evaluated with infants born preterm.
The results of research question 1 support the hypothesis that infants born at term spend more than two-thirds (67%) of the time with their trunks in flexion or neutral. The dominance of flexion or neutral positioning supports the clinical belief that infants born at term do not often position their trunks in extension. This supports the findings of Saint-Anne Dargassies7 that infants born at term tend to position their trunk and extremities in some flexion. Infants born at term are commonly believed to have a dominance of flexion immediately after delivery referred to as “physiological flexion.” Because the infants born at term were assessed at a mean age of 14.5 days, their physiological flexion may already have decreased. The amount of time spent by infants born preterm with their trunks in flexion or neutral was not significantly different from two thirds (67%) of the period monitored. Although the results in Figure 2 suggest that infants born preterm spend less time in flexion or neutral than infants born at term, this difference did not reach statistical significance.
The results of research questions 2 and 3 do not support the hypothesis that infants born preterm flex their trunks less and extend their trunks more than infants born at term. The pattern of results in Figure 3, however, reflects the hypothesized trends. Several factors may contribute to the failure to achieve statistical significance. The criteria of two standard deviations (SDs) above the mean sleeping ratio for extension and two SDs below the mean sleeping ratio for flexion were fairly conservative. This means that a wide range of positions were classified as neutral, possibly making it more difficult to detect group differences in flexion or extension. Perhaps less stringent criteria for categorization of trunk extension and flexion should be used in future studies. Utilizing the infants’ mean sleeping pressure ratio as the basis for determining the neutral position may have produced a bias toward detecting flexion over extension for both groups. During the sleeping segment of the assessment, infants were relaxed on the support surface in a relatively extended position. To meet the criteria for an extended trunk position, an infant needed to be more extended than he or she was in a sleep state. Because the baseline trunk position was relatively extended, further extension may have been more difficult for the infants to achieve than movement into trunk flexion.
Another possible explanation for the lack of between-group differences in flexion is that infants who were very flexed occasionally rolled from supine into side lying, resulting in exclusion of the data from these time periods. Only one preterm infant rolled into side lying during the awake segment. Three infants born at term rolled, each rolling multiple times during their awake segments. These three infants born at term may have been in a flexed position more often than was recorded because of their inability to remain in supine while very flexed.
The hypothesis that infants born preterm would be unsuccessful in their attempts to maintain their trunks in flexion, thus having shorter events of trunk flexion than infants born at term, also was not supported. However, the data depicted in Figure 4 illustrate the anticipated trend. Like the total duration of flexion, the flexion event duration may have been impacted by the more frequent rolling by infants born at term compared to infants born preterm. Each roll shortened what might have been a long flexion event into two short flexion events. The inclusion of the short flexion events may have reduced the mean flexion event duration for infants that rolled. The mean event duration for infants born at term was longer than the maximum event duration for the infants born preterm. The lack of significance may be attributed to the large variability in event duration for the infants born at term.
The large variability seen in the term group may represent the large and adaptable motor repertoire of healthy infants born at term, while the infants born preterm in this study had less variability and possibly a smaller repertoire of more stereotypical movement.12 Another factor that may have increased variability was state stability. Parents of infants born at term often had difficulty predicting their infant’s sleep or awake periods. Frequently, assessments began only a few minutes after the infant had fallen asleep or woken up, making it difficult to capture both awake and asleep states in a single test session. Thus, the quality and consistency of the infant’s state during the assessment may have been reduced. Parents of infants born preterm, on the other hand, often reported their infants were on defined schedules, possibly because consistent scheduling in the hospital was carried over to the home. Infants born preterm had also been home from the hospital for a longer period of time than most infants born at term at the time of the assessment. Therefore, infants born preterm may have had more time to develop a schedule and for their parents to adjust to having a new baby at home. These factors may have contributed to increased variability in the data for infants born at term.
Contrary to the Vles et al study11 of posture and age, gestational age ranges of the infants included in this study may have had a significant effect on the results. The term group had a mean age of 38.9 weeks of gestation. The preterm group included infants ranging from 25 weeks of gestation to 34.6 weeks of gestation with a mean of 31.9 weeks. The three infants in this study who were born preterm at less 30 weeks positioned their trunks in extension for a larger percent of the time and in flexion or neutral for a shorter percent of the time than the 14 infants born at 30 or more weeks of gestation. The inclusion of infants born at older gestational ages may have decreased the number of significant findings. In the future it would be appropriate to include a preterm group with a mean age of less than 30 weeks of gestation.
Infants who were the product of multiple gestations were included in both study samples. However, there were more multiple gestations in the group of infants born prematurely. The intrauterine and extra-uterine environments may differ between singletons and multiple births. Similarly, infants who were born via vaginal delivery may have had different experiences than those born via cesarean sections, especially if the cesarean section was the result of breech or similar intrauterine positioning. Although the small sample size limited our ability to conduct statistical analyses of these factors, they should be considered in the design of future studies.
Beyond the explanations based on methodological issues, there is a possibility that infants born preterm and at term do not differ significantly in the amount of time they spend with their trunks flexed or extended. Perhaps infants born preterm and at term differ in their trunk positions immediately after birth, but this difference becomes negligible by several weeks after birth. Cioni,12 reported a decrease in flexion posture in infants born at term between the first and fourth day of life. Clinically observed differences in flexion and extension may be the result of variations in extremity posture, similar to those reported by Saint-Anne Dargassies7 and Casear,8 without differences in trunk position. Scapular retraction,4 trunk hyperextension2 and abnormal sequencing of flexor and extensor musculature15 seen in some infants born preterm may develop as a result of trunk positions which may not be evident at 41 to 43 weeks gestational age. Perhaps healthy infants born preterm do not position their trunks differently than infants born at term, but rather it is infants with medical complications who account for the variations seen in later development. Longitudinal assessments are needed to determine whether infants born preterm who exhibit excessive extension of their trunks shortly after term age go on to develop motor impairments.
The basic methodology developed for this study may provide a non-invasive method of quantifying infants’ trunk positions during spontaneous activity. Infants were assessed in their natural surroundings while engaged in spontaneous activity, preventing undue stress that may result from muscle tone and motor function assessments that require handling of the infant. The method permitted quantitative measurement of trunk position during spontaneous activity as opposed to the subjective descriptions of trunk position or muscle tone during elicited responses that previously have been reported in the literature. While the results of research questions 2 - 4 were not statistically significant, the trends in the data were in the hypothesized direction, providing some support for the construct validity of the measurement. The reliability of the instrument and the methods were supported during this study. The lower test-retest reliability of this unique methodology should improve with decreased variability in the state of infants being assessed and longer time periods for assessment. In addition, separate researchers should be responsible for collecting and reducing the data. Separation of these tasks would allow for blinding of the researcher who performs data reduction. Blinding eliminates researcher bias that may be introduced when a single researcher conducts the assessment and data reduction as was the case in this study. With further refinement, this methodology may prove useful in documenting trunk positions of infants born preterm and at term in supine, using various positioning devices, and during developmental supportive care in the neonatal intensive care unit. Further studies are needed to compare the use of pressure mapping instrumentation with observation by skilled clinicians.
The results of this study can be used to question the common clinical observation that infants born preterm position their trunks in more extension and less flexion while awake and active than infants born at term. However, these results may be the result of large within-group variability and should be interpreted with caution. With further refinement, the methodology used to assess infants in this study may be useful for quantitative assessment of trunk position in infants.
We would like to thank Vista Medical for loaning us the Ultra Thin Seat Mat to complete this project. We also express our appreciation to the nurses, physicians, and nurse practioners who helped identify infants and families who might be interested in participating in this study. This study would not have been possible without the infants and parents who willingly participated at a very hectic time in their lives.
© 2005 Lippincott Williams & Wilkins, Inc.