The Ethics Committee at the Aragon Health Sciences Institute approved recruiting a cohort for this study. Pediatricians in Section III of the Aragon Health Services referred 34 subjects aged less than 28 weeks having signs of positional plagiocephaly. The inclusion criterion was a difference of at least 5 mm between cranial diagonal diameters , that is infants with moderate or severe PP . We excluded infants that had received other orthotic treatment or physiotherapy or that presented genetic, communicable, metabolic or neurological illness or craniosynostosis.
For the calculation of the sample size, we used non-published data from a previous pilot study with 41 subjects with similar characteristics and receiving a manual therapy intervention similar to the one used in the present study. A decrease of 4,52 ± 2,91% in the cranial vault asymmetry index was obtained in this pilot study. The sample size was calculated using the GRANMO calculator (https://www.imim.cat/ofertadeserveis/software-public/granmo/), with the selection of two independent population means, bilateral contrast, with a α risk of 0.05, a ß risk of 0.20 and a ratio of 1 of the number of subjects between the groups. A minimal number of 7 subjects per group was obtained.
The subjects were randomized into 2 groups with a final number of 17 subjects per group. The subjects were randomized following a design generated with the on-line computer application at www.random.org/sequences. The evaluators were not told about this design.
An informative document about the study was provided to the parents and an informed consent was signed after they had read the document and their questions about the study had been answered. Regulations and guidelines regarding freedom, absence of coercion, disclosure of economic interests, understandable and complete information, confidentiality and acceptance were followed .
The Ethics Committee at the Aragon Health Sciences Institute approved the study (Registry No. C.P. - C.I. PI16/0275. Date: October 25, 2017). The study is registered at clinicaltrials.gov, with identification number NCT03659032. Registration date: September 1, 2018
Clinical and demographic data were extracted from the medical history and the testimony of the parents: age (weeks), birth weight (gr), sex, prematurity, instrumental delivery, firstborn, multiple birth, head positional preference, pediatrician diagnosis of congenital torticollis, plagiocephaly side, transport type and time that the infant spent in prone position awake and watched with 1 month (min) and with 2 months (min).
The following anthropometric parameters, constituting the dependent study variables, were measured: maximum cranial circumference (MCC) , maximal cranial length, maximal cranial width and diagonal cranial diameter taken from the frontozygomatic suture (fz) to the fellow lambdoid suture (lb) . From these data, we calculated the cranial vault asymmetry (CVA) (the difference between the diagonal diameters) , cranial index (CI) and cranial vault asymmetry index (CVAI). CI was calculated by dividing the cranial width by cranial length “Cranial Width ÷ Cranial Length × 100”, while CVAI was calculated using the formula “difference between diagonal cranial diameters ÷ short diagonal cranial diameter × 100” .
Normal range described for CI is between 75 and 85% . According to Mortenson & Steinbok, CVA can be classified into the following categories: normal CVA < 3 mm, mild / moderate CVA ≤ 12 mm, moderate / severe CVA > 12 mm . Plagiocephaly severity scale, according to the CVAI, pursuant to the Children's Healthcare of Atlanta, 2015  is: level 1: <3.5%; level 2: 3.5 to 6.25%; level 3: 6.25 to 8.75%; level 4: 8.75 to 11.0%; level 5:> 11.0%.
At the end of the study, the parents were given a visual analogue scale (VAS) to evaluate their perception of head shape change . The parents made a vertical mark on a line graduated from − 10 (much worse) to + 10 (much better); there was an intermediate Item 0 (no change).
17 subjects received 10 sessions of manual therapy and a caregiver education program, an integrative concept of treatment that will be identified in the manuscript as pediatric integrative manual therapy (PIMT). Each PIMT treatment for remodeling the cranial deformation consisted of:
one maneuver to mobilize the neuromeningeal tissue at the lumbosacral level, based on the technique of John E. Upledger  but adapted to the pediatric field. Very light traction is induced through the pelvis to stimulate a tissue response. The physiotherapist follows the movements of the baby's pelvis according to the active movements, trying to move towards increasingly flexed positions in order to act on the general dural tension.
one technique for the cervical spine based on the works of Giammatteo . Very slight traction is applied through the head and the active movement of the head is accompanied to different positions of flexion and extension, lateralization and rotation, stabilizing the atlas gently in a dorsal direction.
one technique applying manual pressure to mold the skull base in the opposite direction from the PP torsion on the skull base, based on the work of Arbuckle . The manual pressure was applied to the occipital bone to displace it dorsally, insisting on the flattest area, representing attempted cranial torsion (Fig. 1).
This PIMT protocol was applied by several pediatric physical therapists with specialized training and 4 years of experience. The results obtained with PIMT treatment in cervical spine mobility are described in another manuscript pending publication.
Each manual therapy session lasted 20 minutes, once a week.
The caregiver education program consisted of a series of literature-based recommendations [52, 53] that encompassed repositioning, sensory and motor stimulation of the opposite side to the prefered one and prone positions. Parents were instructed with the help of a trained pediatric physiotherapist and an informative booklet about basic recommendations.
The 17 subjects from the control group received solely the same caregiver education program. The control group was convened once during the 10 weeks to control their evolution, listen to their difficulties, solve their questions and insist on the importance of the program of stimulation and positional advice.
The Kolmogorov-Smirnov test with the Lilliefors correction was used to test the normality of the distribution of the quantitative variables; the Shapiro-Wilk test was used for this purpose if n < 30. A descriptive analysis of the qualitative variables was carried out, offering the absolute and relative frequencies; as well as a descriptive analysis of the quantitative variables, offering the mean ± standard deviation or the median (Q1; Q3) values, depending on whether the distribution of the variables was normal or not, respectively.
If distribution was normal, the Student t-test for independent samples was used for the intergroup comparisons of the dependent pre-intervention variables. The Mann-Whitney U test was used for these comparisons if the distribution was not normal. To compare intervention effectiveness between the groups, we calculated the improvement indexes of the dependent variables using the difference of the final measurement values minus the baseline measurement values. If the distribution was normal, the improvement indexes were compared using the Student t-test for independent samples; if not, the Mann-Whitney U test was used. The effect size of the interventions evaluated was calculated using Cohen's d.
We analyzed the correlations between the improvement index of the CVAI and of the VAS reflecting the parental perception of head shape change in the entire sample. To do so, the Pearson correlation coefficient was calculated when the variables followed a normal distribution. If the distribution of a variable was not normal, the Spearman Rho coefficient was used.
A confidence interval of 95% was established for the analyses. Statistical significance was set at p < 0.05. The statistical study was performed following the principles of intention-to-treat analysis, without attributing values in the second assessment to the subjects lost to study throughout the intervention.