General Movements in preterm infants undergoing craniosacral therapy: a randomised controlled pilot-trial

Preterm infants, with a gestational age (GA) [21] between 25 and 33 weeks who were admitted to the neonatal intensive care unit of the University Hospital of Graz, Austria, were eligible. As we planned a treatment period of 3 weeks, and most infants are discharged with a GA of 37 weeks, the upper limit of inclusion was 33 weeks. Infants with congenital anomalies, presence of major malformations, any abnormality in cranial ultrasound, elevated bilirubin levels, and any need of respiratory support (need for oxygen or mechanical ventilation) during the study period were excluded. To guarantee that only preterm infants with subsequent normal neurodevelopment were included, follow up was done regularly at the corrected age [21] of 12 and 24 months. The neurological examination was performed according to Touwen [22], for neurodevelopmental testing the Bayley Scales of Infant Development were applied [23]. Some infants that were initially included were excluded later on due to signs of abnormal neurological development. Thus the inclusion of only neurological healthy infants was guaranteed. The ethics committee of the Medical University of Graz approved the study and written parental consent was obtained prior inclusion (20-009ex08/09). The trial was registered with the German Clinical Trials Register DRKS00004258.

Newborns were assigned to the intervention group (IG) or to the control group (CG) in a 1:1 ratio using a randomised block design with block size of 6. A sequentially numbered, sealed, opaque envelope containing the allocation was opened by a researcher after parental consent. Infants randomized to the IG received a total of six interventions. Interventions were standardised to 20 min/treatment with a frequency of two intervention/ week over three weeks. The CG did not receive the standardised intervention or any other kind of OMT interventions during this period (Fig. 1).

The intervention was performed by two specialized physiotherapists trained at the Upledger Institute Graz, Austria, with 10 years of experience and certified as Upledger CranioSacral® Therapists. The two physiotherapists involved in the study were trained to use only indirect and fluidic techniques. None of the physiotherapists were involved in the study design, data entry or statistical analysis. Both groups received usual care, which means – traditional (~routine) nursery care e.g. side lying and “nesting” in an incubator and skin to skin contact by the parents. In addition all parents, the physicians and the GM assessors were unaware of patient allocation except for the treating physiotherapist.

The intervention was scheduled according to the circadian rhythm of the infant (e.g. intervention was only started when the infant was awake and after feeding to avoid influence of agitation due to hunger). Infants were positioned in supine position, naked, wearing only swaddling bands. After initial contact with the child, by using the preferred initial touch, based on the requirements of basal stimulation, the specialized physiotherapist started with the evaluation of the craniosacral system according to a (modified) 10 step-program [7, 24]. The 10 step-program was modified as follows: exploration of the cranial system (step 1), treatment of asymmetry (step 2), evaluation of the overlapping of the cranial bones (step 4), exploration of the balance of the membranes of the cranial and spinal dura mater (step 7), exploration and treatment of the sacrum (step 8), and exploration and treatment of the chest (step 9). After the evaluation craniosacral therapy was initiated to achieve the greatest relaxation.

The preterm infants were videotaped twice a week: in the IG 5 min before and after each intervention (which means before and after OMT), and same in the CG but without intervention.

The videos (10 min per session) were taken with the infants lying supine in the incubator wearing only diapers. Each infant was recorded 12 times, which corresponds to a 2-h footage per infant.

As the GMA is dependent on the behavioural state [26] of the infant, only infants in behavioural states 2 (i.e., eyes closed, irregular respiration, general movements present) or 4 (i.e., eyes open, irregular respiration, general movements present) were included. If an infant was in behavioural states 1, 3, or 5 (corresponding to quiet sleep, quiet wakefulness or crying), his/her GMs could not be assessed. Therefore, the number of infants varies slightly within the weekly assessments. All videos were edited according to the standards of GMA [27]. Normally, the GMs of a preterm infant comprise the entire body and manifest themselves in a variable sequence of arm, leg, neck and trunk movements. They appear and cease gradually, varying in intensity and speed. Rotations and frequent slight variations of the direction of motion make them look complex but smooth [15‐18]. GMs are categorized as normal (N) or abnormal. Abnormal GMs are classified into (1) “poor repertoire GMs” (PR), whereby the sequence of movement components is monotonous; the amplitude, speed, and intensity lack the normal variability; (2) “cramped-synchronized GMs” (CS), which appear rigid as they lack the usual smoothness and fluent character; the limb and trunk muscles contract almost simultaneously and relax almost simultaneously; and (3) “chaotic GMs” (Ch), which appear jerky and abrupt due to their large amplitude and high speed [15, 18].

Besides the global GMA into the categories N, CS, PR, Ch (primary outcome) mentioned above, we assessed as detailed GMA, the General Movement Optimality Score (GMOS) as secondary outcome based on Prechtl’s optimality concept [17, 28]. The detailed GMA (GMOS) is a further tool to assess preterm motor movements. Thus it is a supplemental tool to define motor optimality with the principles of GMA. The GMOS is composed of the subscore for the sequence of movements (maximum = 2) and three sub-component optimality scores: (1) optimality for neck and trunk movements (maximum = 4); (2) optimality for the upper limb movements (maximum = 18); and (3) optimality for lower limb movements (maximum = 18). A higher score indicates a more optimal performance. For the GMOS, the maximum composite score of 42 indicates the most optimal GM performance. All GMs judgements were done by two experts in GMA (P.B.M., C.E.), who were blinded to group assignment, after completion of the 3-week structured program. In case of disagreement on particular details the recordings in question were re-evaluated until consensus on the final score was achieved. GMA has been validated as a diagnostic tool for detecting early brain dysfunction in newborn infants [19, 29‐31]. The early identification of individual infants at high risk remains difficult. Spittle et al reviewed the clinometric properties of neuromotor assessments for preterm during the first year of life [16]. Furthermore, Bosanquet et al , Burger et al and Noble et al showed that Prechtl’s assessment of the quality of GMs offers the best combination of reliability, sensitivity and specificity for predicting cerebral palsy in the early months [19, 32, 33].

To compare global GMA (N vs. PR vs. CS) between groups before the first session and after the last session Fisher’s exact test was used.

To analyze GMOS a linear mixed model with fixed effects for session, time point (time point refers to the comparisons of the measurements before vs. after each session) and intervention (IG vs. CG), random effect for individual children and a first order autoregressive covariance structure was used for calculation of significant differences between groups and interactions. Following interaction terms were included in the model: session*time point, session*intervention, time point*intervention and session*time point*intervention. Exploratory post hoc analyses (linear mixed model, fixed effects: session, time point and intervention, random effect: individual children, interaction: session*time point*intervention) were performed to determine group differences for all twelve measurement (before and after all six sessions) separately. To compare patient’s characteristics Pearson Chi Square test and Mann Whitney u-Test was used. To analyse inter-scorer agreement Cohen’s Kappa was calculated. A p-value of a p < 0.05 was considered statistically significant. Statistical analysis was performed using IBM SPSS Statistics 22 (SPSS Inc. Chicago, IL, USA, 2013).

Of 300 video recordings, 28 (9 %) had to be excluded, because the behavioural state of the infants did not allow proper assessment, and a further 4 (1 %) were excluded due to technical problems during recording. The footage for detailed analysis consisted of 268 assessable recordings and a total of 66 h recording time. The global GMA was done off-line from video and usually took the experienced observer 3-5 min per recording (Mean: 4 min; total 17.5 h of analysis). The detailed assessment (GMOS) was done offline from some video and lasted between 20 and 30 min per recording (Mean: 23 min; total 112 h of analysis).

Primary outcome: GMA

The GMA at the beginning of the interventional period showed that 5 infants had normal GMs (N) and 20 a poor repertoire of GMs (PR). There was no difference in global GMA between the two groups (IG: 1 N, 11 PR; CG: 4 N, 9 PR; Fisher’s exact Test P > 0.05). At the end of the interventional period GMA were again comparable between groups (IG: 2 N, 9 PR, 1 CS; CG: 5 N, 7 PR, 1 CS; Fisher’s exact Test P > 0.05). In each group one infant with a PR of GMs improved to N and one infant deteriorated from PR to CS.

Secondary outcome: GMOS

GMOS did not change from session to session (main effect session: p = 0.262). Furthermore no differences between IG and CG (main effect group: p = 0.361) and no interaction of session*group could be observed (interaction session*group: p = 0.658) (Table 2). Post hoc analysis showed a trend toward higher values before (post hoc analysis for group differences at session 1, and time point 1: p = 0.085) and after (post hoc analysis for group differences at session 1, and time point 2: p = 0.075) the first session in CG compared to IG group. At all other time points GMOS were not significantly different between groups. (Table 3, Fig. 2).

For the GMA (primary outcome) the inter-scorer agreement revealed a Cohen’s Kappa of 0.87; for the GMOS (secondary outcome) Cohen’s Kappa was 0.76.

We would like to emphasize that i) only infants with normal and adequate neurodevelopmental evaluation remained within the study groups; therefore, all included infants (in both groups) showed adequate normal neurodevelopment at the age of two years, ii) this was a randomised trial, iii) we used an established assessment instrument (general (primary outcome) and detailed GMA (secondary outcome) and iv) the assessors of the GMs were completely blinded with respect to group assignment. In addition, there was no difference in terms of their diseases of prematurity, the diagnosis of respiratory distress or bronchopulmonary dysplasia [53] or the duration of mechanical ventilation.

We only included a small sample size, which is a limitation of the study. We planned enrollment of 12 subjects per group, as proposed by Julious [54] and Billingham et al [55]. An additional 20 % (total of 15) was included for each treatment arm, anticipating subject withdrawal or other unforeseen postenrollment exclusions from the study [54, 55]. Based on our results we conducted power analysis and sample size calculation. Our study has a power of 22 % to detect an interaction of intervention*time. In a future trial using the same approach, 57 children in each group would have to be included to detect a significant treatment effect (interaction of intervention*time) with a power of 80 %.

The following questions arise: Are two interventions per week enough to see any significant effects? Would the results be different by using another kind of OMT? Should further studies include a control group with “touch” and “presence” of the therapist?

Nevertheless, the results of the present study underline the fact, that using craniosacral therapy in preterm infants is safe, as there was no deterioration of neurodevelopmental results in IG.