J Manipulative Physiol Ther 2015 (Nov); 38 (9): 713–726 ~ FULL TEXT
Aurélie M. Marchand, DC, MScACPP
Doctor of Chiropractic,
OBJECTIVE: A literature review of tensile strength of adults and pediatric human spine specimens was performed to gather information about biomechanical forces and spinal differences of adults and children and to synthesize these findings into a scaling model to guide safety and clinical decisions for spinal manipulative therapy (SMT) for children and infants.
METHODS: The literature search was performed using PubMed from inception to November 2012 with no filters or language restrictions. The search included terms related to pediatric spine biomechanics and tensile strength. Studies included those in which human tensile strengths necessary to create anatomical damage in the cervical, thoracic, or lumbar spine of pediatric human subjects were recorded. The pediatric population was defined as human subjects from birth to 18 years old. Biomechanical findings were used to propose a scaling model based on specimen age and differences in tensile strengths. A model of care was proposed using the scaling model and the existing literature on pediatric technique adaptations.
RESULTS: Nine experimental studies were selected, 5 in the pediatric population (46 specimens) and 4 in the adult population (47 specimens). Mean tensile strengths were estimated, and ratios were used to describe differences between 4 age groups. The preliminary model of care proposed includes maximum loading forces by age group. From these studies, a model showed a nonlinear increase in the cervical spine tensile strengths based on specimen age.
CONCLUSIONS: The literature showed that tensile strength differences have been observed between pediatric and adult specimens. A preliminary model of care including pediatric SMT technique adaptation based on patient age is proposed, which may possibly contribute to further knowledge of safety and clinical implications for SMT for children and infants.
Key Indexing Terms: Chiropractic, Children, Therapeutics, Manipulation, Spinal
From the Full-Text Article:
Spinal manipulative therapy (SMT) is currently used to treat and manage a wide variety of musculoskeletal conditions.  The application of precisely controlled high-velocity, low-amplitude thrust to a joint during SMT causes tissue deformation of the spine and surrounding tissue.  The thrust is designed to restore motion in the targeted joint by applying force to the area of segmental restricted motion. [2, 3] Theoretically, the thrust is applied in the paraphysiologic space of joint motion  while taking care not to exceed the anatomical limit leading to joint trauma and pathology. The osteokinematic movements and arthrokinematic movements have guided the rationale and the application of SMT in adult patients by considering the mechanical forces introduced in a joint (tension, compression, shear, torque) in relation to tissue properties contributing to kinetic joint stability and integrity (muscles, ligaments, facet joints, intervertebral disks).  Spinal manipulative therapy and thrust application are based on aspects including anatomy, tissue properties, and spinal biomechanics, as is noted by the lower loads applied in the cervical spine compared with the stiffer thoracic spine. 
Chiropractic care of children is common, and it is estimated that between 5% and 20.5% of chiropractic patients are of pediatric age. [5–10] It is generally agreed that manual practitioners should adapt SMT techniques depending on the patient's size, structural development, flexibility, and preference.  Because pediatric spinal biomechanics differ from adults because of anatomical differences in tissue tensile properties (stemming from size and tissue organization variations),  it is logical that these factors are considered when addressing biomechanical conditions for various age groups. European chiropractic practitioners have reported technique modifications including adaptations of speed and force during the application of spinal manipulation and differences in obtaining cavitation sounds in 5 pediatric age groups ; however, these were observational only. Currently, technique adaptations such as the amount of force to use for SMT on infants and children are supported by little evidence. [11, 13–15]
The safety of chiropractic treatment for young infants (1–12 months old) was raised in a study observing vegetative responses (bradycardia and apneoa) after a thrust ranging from 30 to 70 N of force (average of 50 N) applied to the upper cervical spine.  The study suggested that in the prone sleeping position, “the possibility that a minor mechanical irritation of the cervical region may trigger the first step in the events that lead to SID,” or sudden infant death syndrome, and “the chiropractic impulse that triggers a bradycardia and apnea suggests that comparable mechanical stimuli associated with the prone position may result in similar adverse responses.”  In other words, this study suggested that death secondary to sudden infant death syndrome might occur after chiropractic SMT of the cervical spine in pediatric patients younger than 1 year. On the other hand, the same study reported that 20,000 children were treated with chiropractic manipulation without reports of serious adverse reactions,  while a literature review reported 9 cases of serious adverse reactions in the pediatric population (from 1900 to 2004), of which 2 cases were not attributed to chiropractors.  Other studies reported adverse effects and nonserious adverse reactions secondary to pediatric SMT, with occurrences ranging from 0.23% to 9% of the pediatric population. [13, 17–19] Because the application of force and responses have been linked,  it is important to consider how much force should be used in SMT for the pediatric population to inform practitioners and decrease the likelihood of adverse reactions. Scaling models of the adult and pediatric cervical spine currently exist,  and 2 studies reported pediatric SMT technique adaptations in patients. [11, 13] To the author's knowledge, scaling methods have so far not been applied to the clinical practice of pediatric SMT.
Because it would not be appropriate to conduct a trial testing SMT that would cause tissue failure in children and infants and to address the gap of knowledge of SMT force and children, it was decided to conduct a literature review. The purpose was to identify the amount of force necessary to create damage in the pediatric spine, which could be used as a limit of force that should never be exceeded during pediatric SMT. Adult data were identified to compare with pediatric data to evaluate differences between pediatric and adult spines so that scaling models could be proposed. A model of care is discussed to address the gap of knowledge concerning pediatric SMT technique adaptations to prevent the occurrence of safety incidents.
Although a previous literature review has been published on tensile strength, none have been published recently.  This review aimed to identify the differences between adult and pediatric spines from the clinical aspect of SMT. The results showed a nonlinear increase in the cervical tensile strength in relation to increasing age of the specimens, as previously described in the literature. [12, 22, 23, 26] The tensile properties of the human adult and pediatric spine described in the results originated from 9 articles, of which 1 reported both pediatric and adult values, 4 were adult studies (a total of 55 adult specimens, 47 included), and 4 were pediatric (total of 61 pediatric specimens, 46 included) studies. These articles helped fill a gap in the literature arising from the necessity of the automobile industry to increase the safety of pediatric passengers. Eight articles provided cervical spine data on the basis that cervical spinal trauma is most frequent and may potentially lead to lifelong sequelae in comparison with thoracic or lumbar trauma. [12, 37] The 9 experimental studies on human tissues form part of the literature along with scaling techniques (from animal studies) and computer models that aimed at producing accurate estimation and prediction models of tissue damage. A better understanding of the biomechanics of the pediatric spine in response to stress during a car crash has helped the automobile industry prevent trauma in both adult and pediatric populations. This review may be similar in its intent because it aims at providing a better understanding of the differences between the adult and the pediatric spine to help decrease the risk of serious adverse reactions after the incorrect use of pediatric SMT.
The selected studies were not randomized controlled trials, which led to difficulties in assessing risk of bias according to the standards established by the Cochrane collaboration tool.  It was decided to review the limitations reported in the selected studies. The limitations reported were similar with regard to the use of cadaveric specimens, small sample sizes, and exclusion of tissues (skin, musculature). It was observed that some studies did not report the location (center of gravity or vertex) of axial loading. As indicated by Dibb et al,  the location of traction leads to significant differences in tensile strength. The omission of this information may create a bias in correctly estimating the tensile strength of the cervical spine and may have created an error in the synthesis of collected data proposed in this review. For this reason, the mean tensile strength across studies was reported, whereas no attempt was made to obtain the minimum, maximum, or standard deviations (nonhomogeneous data).
The results proposed by this review could be extrapolated to the live human population; for this reason, the exclusion criteria were strict with regard to animal studies and computer models. The results were obtained from cadaveric specimens that may be representative of the population because the specimens ranged by stature and weight from below the 5th percentile to the 50th percentile and above.  The reported differences in tensile strength between studies may be related to differences in specimen size.  Special attention was paid to comparing similar intervention (tensile strength) and similar human tissues (osteoligamentous spine and intact specimens). Cervical spine ratios were obtained from the osteoligamentous spine only. There were variations in the methodology described to obtain the tensile strengths in the selected studies. For studies reporting axial loading along the center of gravity or along the vertex, it was decided to report both values (Table 1) and obtain the mean value for the different methodologies.
Because the methodology of axial traction was not defined in 2 pediatric studies, it was decided that obtaining the mean tensile strength would be appropriate. A great difference in methodology was observed in a study from 1874 ; however, these results were excluded for ratio calculations. The tensile strength values selected for inclusion were based on the lower thresholds reported. Because, clinically, the lower values would be more relevant to prevent tissue damage compared with ultimate failure loads, the initial failure load was selected in 2 studies. [22, 23] Similarly, the major failure load was selected instead of the ultimate failure load to allow for better comparability.  For the purposes of this work, the cervical spine was considered as a whole rather than being divided into the upper and lower areas. Despite the difference in mean tensile strength between the upper and the lower cervical spine, it was arbitrarily decided to group the cervical spine as one region in an attempt to provide more meaningful results based on a larger number of specimens. However, there are 3 main reasons for which the upper and lower cervical spine means could be considered different: the differences in tensile strengths between those segments, the pathomechanical properties of the cervical spine (the fulcrum of motion moving from upper to lower cervical with increasing age),  and increased laxity of the occipitoatlantoaxial complex.  The values from the upper and lower thoracic spine data were obtained from 4 specimens; it was considered relevant to also group the thoracic spine as one region.
Only the values obtained from the failure loads in the osteoligamentous spine were considered to obtain ratios because full spine specimens had higher values owing to the increased resistance offered by soft tissues, as previously described in the literature. [23, 26, 30] It was arbitrarily decided to use the osteoligamentous spine because more specimens were available for comparison. Differences in spinal tensile strengths may be caused by specimen preparation, with higher values observed when full cervical spine including muscles, ligaments, skin, and fat are used, [22, 27, 29] compared with studies on osteoligamentous spinal tissues only. [26, 29] Possibly muscles may play a role in increasing the tensile strength of the pediatric spine. In adults, maximal muscular activation increased cervical tensile tolerance from an average of 1800 N (osteoligamentous complex) to 4800 N (maximal muscle activation).  On the other hand, muscles have been reported weaker in children and muscle attachment sites (spinous processes) are smaller in immature spines,  thereby decreasing the efficiency of muscle contractions in children compared with adults. The tensile strength of the intact adult cervical spine was 3174.0 N  and that of the intact newborn was 507.2 N,  for a ratio of 0.16. However, the methodology used by Duncan  could be considered crude (whole spinal column) and was not comparable to that of Yoganandan et al  (cervical spine). The mean tensile strength across studies was reported. No attempt was made to obtain the minimum, maximum, or SDs.
The data on premature specimens were not included for analysis because the age range of the premature specimens was from 20 to 37.5 weeks' gestation [22, 23] and may have skewed the result for full-term infants because the values were lower compared with full-term newborns. In the study of Duncan,  the age of 4 specimens was not specified and referred to as “adult female/male foetus.” These values were included in the calculations to obtain the average spinal column tensile strength but were not included in the calculation of the ratios.
The age groups were created to separate and regroup children on the bases of 6 main factors:
the pathomechanical properties (fulcrum of motion), 
mechanical properties including bone density changes, 
anatomical differences, 
head ratio changes, 
muscle strength, [22, 39]
and cervical axes of motion. 
This literature review aimed at addressing a gap of knowledge in the technique adaptations used during pediatric SMT based on biomechanical properties of the tissues involved. Information from the studies found in the review was combined into a table so that a preliminary model could be constructed. The ratios proposed in Table 2 were obtained from the mean tensile strengths of the cervical spine and divided by age groups. The results from the various studies, when combined together, suggest that there is a nonlinear increase in the tensile strength of the cervical spine based on increasing age.
During SMT of the adult spine, thrust is commonly applied in the paraphysiologic space of joint motion  while taking care not to exceed the anatomical limit, leading to joint trauma and pathology. For the development of this model, the tensile strength was considered to be the extreme end of the anatomical limit where tissue damage occurred; this limit was also referred to as the catastrophic level.  Therefore, the extreme limit at which pediatric SMT could be applied would be reflected by the tensile strength of the tissues involved, in this case the osteoligamentous spine. This assumption was based on the observations that tensile strengths increased with increasing subject age [22, 23] and that technique adaptations reported in pediatric SMT also increased (force, speed) with pediatric age. [11, 13] Indeed the consensus of Hawk et al [11, 13] seemed to indicate a progressive increase in the forces used during pediatric SMT (age being an important factor in patient size, structural development, flexibility) along with European chiropractors reported technique adaptations progressively changing according to the age of pediatric patients.  Both these clinical aspects seemed to be reflecting the biomechanical findings of increased tensile strength in relation to specimen age.
Safety Implications of Model Proposal
Adult SMT occurs in the paraphysiologic zone of joint range of motion.  Values obtained from the tensile strengths of the cervical spine may be considered the extreme anatomical barrier where tissue damage occurs after the application of a load. During the initial phase of model development, it was proposed to not exceed forces of 197 N in the cervical neonatal spine because such load has been shown to lead to tissue damage. Similarly, it was proposed to avoid using of loads above 560 N in the cervical spine of 1– to 23–month-olds, 970 N in 2– to 8–year-olds, 1500 N in 8– to 18–year-olds, and 1750 N in adults to prevent tissue damage, as suggested by the data obtained from this review. However, this approach was not considered relevant and was therefore discarded for the reasons explained below.
Luck et al  reported that models based on solely on the tensile strength of spinal segments may not take into consideration subcatastrophic tissue damage occurring at lower values. Indeed, the displacement observed during mobile testing of the cervical spine shows that the pediatric spine can withstand higher structural displacements (up to 5 times more) compared with adult spine, but the pediatric spinal cord was not able to withstand such displacement without serious compromise (spinal cord injury).  Therefore, the values obtained from the tensile strengths of specimens may not accurately represent the subcatastrophic level at which spinal cord injury may occur. However, the values obtained from the mean cervical spine tensile strength may indicate an absolute upper limit that should never be exceeded during the application of SMT.
The second reason for excluding the values of the cervical tensile strength was that in the cervical spine, neither adult nor pediatric SMT occurs solely in the direction of axial loading (direction in which the data were gathered in tensile strength studies), but rather occurs with a combination of movements that may include lateral flexion, rotation, extension, or flexion of the cervical, depending on the technique used.  In the thoracic spine, the reported methodology of repetitive loading conditions of 50 cycles with increasing amplitude  was unlikely to be comparable with the high-velocity, low-amplitude thrust used during SMT.  Therefore, the data collected did not allow to make assumptions the values being representative of an SMT intervention because it remained unclear whether axial loading strength was higher or lower than rotational, lateral flexion, flexion, or extension strengths of the spine. 
Because subcatastrophic loads were not identified by this review, it was decided to attempt to provide normal loads parameters based on the loads currently reported during adult and pediatric SMT. Indeed, during SMT, the limit of the anatomical barrier should not be reached and loading values used may be used as a reference for therapeutic (nontraumatic) intervention. The proposed ratios were applied to obtain the normal loads that may be safely used during SMT. The loads applied by physiotherapists to C2 on adults averaged 24 N for grade 1 mobilization technique (small amplitude of movement near the start of the range). [46, 47]
In neonates (0–1 month old), pediatric SMT is consistent with nonforce or low-force therapy  and may be similar to a grade 1 mobilization techniques used by physiotherapists. The reported “safe” amount of force used during chiropractic treatment for those younger than 12 weeks is 1 to 2 N with a touch-and-hold technique.  Using the ratios proposed in Table 2 (known value for an adult patient [24 N], ratio of 0.11 to obtain the value for a patient 0–1 month old), 24 N would be multiplied by 0.11 (ratio) to obtain 2.6 N as an indicator/guide of the amount of force that could be used for the treatment of the neonatal cervical spine. The 2.6-N value obtained by the ratio is slightly above the 1– to 2–N range currently reported for the pediatric chiropractic touch-and-hold technique. 
In an attempt to identify the subcatastrophic level, the average value of 50 N of force in infants (1–12 months old) were used in comparison with known SMT values in the adult cervical spine. This 50-N limit was used because this force was able to produce transient adverse effects in 279 (or 40.1%) pediatric patients (total 695), of whom 68.8% has mild bradycardia and 31.2% had severe bradycardia.  Using the ratios proposed in Table 2 (known value for an infant [50 N], ratio of 3.1 to obtain the value for an adult patient), the value of 50 N would be multiplied by 3.1 to obtain 155 N. The forces of cervical manipulation have been measured at 117.7 ± 15.6 N  and 116 N (41-193 N),  and 104 N in healthy adult subjects.  Forces of 155 N would exceed the average 112.6 N of force used during cervical SMT by 43 N; in other words, forces of 155 N would represent 137.7% of a normal load (112.6 N = 100%). The ratios to identify (known value for adult 24 N, ratio of 0.32 to obtain the value for infants 1 to 23 months old), the normal load using a touch-and-hold technique in the cervical spine would indicate a load of 7.7 N to be normal in patients aged between 2 and 23 months.
The therapeutic intervention described by Koch et al [14, 51] was a “short gentle thrust administered to the suboccipital region…,” which is not similar to touch-and-hold therapy; however, there was insufficient information to allocate the type of treatment to a specific grade of mobilization, as defined by Maitland and others. [46, 52, 53] In a further attempt to compare the value reported by Koch et al to SMT, the average value of 112.6 N reported in cervical adult SMT could be multiplied by 0.32 (ratio) for a loading value of 36 N in the 1– to 23–month-old patients. The reported range of force used was from 30 to 70 N,  which would be similar to the 36-N loading obtained by the ratio in the lower range of loading reported but could represent 139% of “normal loading” for 50 N and 194% of “normal loading” for 70 N.
Responses observed after the application of the average load of 50 N in the upper cervical spine decreased after the third month of age.  The discussion stated that “a mild irritation of the cervical region will more likely lead to a severe bradycardia in the first 3 months” ; indeed, 56.6% of patients younger than 3 months had mild bradycardia (43.4% experienced severe episodes), whereas 75.6% of patients older than 3 months had mild bradycardia (24.4% experienced severe episodes). However, this finding does not consider that children gain head control around 3 months of age.  It was shown that muscle activity increased the tensile strength of the cervical spine ; therefore, it was deduced that the same loading of 50 N before and after gain of head control would have different effects. Before 3 months of age, muscular activation would not be an important factor during therapeutic intervention, whereas after 3 months of age, muscular activation could help increase the tensile strength of the cervical spine. Based on the observation that an average of 50 N of force was able to produce vegetative responses in 40% of pediatric patient, this value was used as a parameter to evaluate sub catastrophic level of loading.
The amount of force delivered during manual therapy has been implicated as a cause for side effects and adverse reactions after care.  The reference of 50 N of force may help quantify the limit of the subcatastrophic clinical level and help practitioners remain within the paraphysiologic zone during pediatric SMT and prevent the occurrence of safety incidents related to incorrect technique. The values calculated with the ratio with 50 N could possibly indicate in the cervical spine a maximum loading level of 20 N in neonates, 85 N in 3– to 8–year-olds, 135 N in 8– to 18–year-olds, and 155 N in adults. It is proposed not to exceed these forces during the application of SMT to prevent the possible occurrence of adverse reactions by the incorrect application of technique.
Clinical Implications of Model Proposal
There is a wide occurrence of pediatric SMT in the chiropractic profession as a management option for many conditions.  Clinically, chiropractic pediatric technique adaptations were previously described by a consensus  and by a European survey.  The 2009 consensus reported that patient size, structural development, flexibility, of joint and patient preferences should be taken into consideration for the treatment for infants, children, and adolescents.  The survey of European chiropractors investigated their opinions with regard to statements on force, speed, and cavitation when using pediatric SMT compared with adult SMT.  It was observed that opinions varied greatly between 5 pediatric ages (0–2 and 3–23 months and 2–5, 6–12, and 13–18 years). Between 6 to 18 years old, there were high agreement rates to the statements on speed and cavitation; it may be assumed that SMT was occurring. Indeed, SMT is defined as a high-velocity, low-amplitude manipulation. [2, 4, 50] In younger children, most respondents strongly disagreed with using the same speed as in adults and strongly disagreed with obtaining cavitation from SMT; the use of high-velocity, low-amplitude SMT for those younger than 6 years may therefore be debatable. The proposed grading may help practitioners record the technique adaptations used more accurately. Currently, the therapeutic intervention of SMT in the pediatric population has been described as pediatric SMT,  pediatric manual therapy,  touch-and-hold therapy, and low-force, low-amplitude manipulation. The current terms used to describe treatment in pediatric patients may be misinterpreted by readers with limited knowledge on pediatric chiropractic technique adaptations who might possibly interpret that intervention is similar to adult SMT.
A chiropractic model of care in pediatric patients is proposed in Table 4 based on the opinions of European chiropractors on force and speed  and on the ratios obtained from the tensile strengths of the cervical spine in different pediatric age groups. The grading proposed artificially separated the pediatric population into 4 age groups. It is proposed to gradually increase the amount of speed and force throughout the age groups for the following reasons. During adult and pediatric SMT, the intervention is designed to restore motion in the targeted joint. [2, 3] In the cervical spine, 1 to 2 N of force was reported as a normal load for treatment for patients younger than 12 weeks,  and an average of 112.6 N was reported in the adult population; these findings indicate that an increase of loading occurs from younger to adult patients. The results of this literature review showed a nonlinear increase in the tensile strength of the cervical spine according to specimen age.  As the tensile strength of the cervical spine increases with age, higher loads may be required during therapeutic intervention to create the required tissue deformation aimed at restoring articular motion.  Clinically, the model of care proposed would indicate that in a 5–month-old child, 30% of adult SMT loading could be used. A clinician should consider that the tensile strength of a 5–month-old child is lower than that of a 20–month-old child and should modify his/her technique accordingly and gradually increase the loading starting from the proposed value of lower age grade. In this example, a clinician based on his/her clinical judgment may initially decide to use slightly more than a 10th of adult SMT loading as well as using his/her clinical experience and training for guidance.
It may be noticed that grade 1 was extended to 2 months of age (instead of 1 month in the proposed ratio) to allow for premature children to reach their structural development and to possibly minimize the use of excessive loading in neonates and young infants. Similarly, the proposed technique adaptations are slightly more conservative than the scaling ratio obtained in an attempt to minimize the use of excessive loading in the pediatric patient. Clinically, treatment in the cervical spine before the developmental milestone of head control has been mastered should be considered an important factor in technique adaptations because it was observed that muscle control increased the tensile strength of the cervical spine  and that the same amount of loading was linked to higher occurrence of vegetative responses in children younger than 3 months. 
It is beyond the scope of this article to cover the anatomical differences of the pediatric spine and the differences in cervical axes of motion and active range of motion, which have been previously been addressed in the literature. However, these aspects should also be considered by the clinicians during pediatric SMT. When joint mobility is deficient in 1 or more plane of motion, clinicians should be able (according to their training and experience) to detect and address the plane of deficit. Until further mechanical plane-specific research is performed, it will be difficult to propose ratios in relation to a specific plane of deficit. The proposed model of care may be used conservatively for all areas of the body and for all planes of motion deficit until further research is conducted. Thoracic ratios may be used at the discretion of the clinicians, considering that those were based on only 2 pediatric human specimens. At all times, clinicians should use their discretion in their applications, accommodate for patient preferences, and follow the first principle of “Do no harm.”
The decision of applying a specific grade, mobilization, or manipulation should be based on patient age, weight, height, sex, neurologic development, muscular control, patient preferences, clinician confidence, and experience, with the aim of restoring function to the articular segment affected. The age groups proposed in this preliminary model of care are based on averages of physical development in the specimens from the selected studies. In the model, the selection of grades 1 to 4 and their related technique adaptations should be tailor made to the patient because each child is individual in its physical development. Pediatric SMT should only be performed by qualified clinicians from a professional background, guaranteeing skills and safety during manipulation and/or adjustment. This would include professions with an accredited curriculum teaching the subject of spinal manipulation along with pediatric SMT training.
First, this is a proposed model using data from previously published studies. More research will be needed to confirm the proposed model; thus, this research should be considered preliminary. The search terms and inclusion and exclusion criteria are limitations to this study. Animal studies were excluded from the data-gathering process because there is a disputable and not well-understood correlation between animal model and pediatric human spines.  On the other hand, the tensile strength of juvenile animal surrogates estimated human cervical tissue stiffness fairly well, although the strength may not have been accurately estimated from these models,  and there is congruency of data based on previous vertebral geometry studies [55–57] and on tensile stiffness of isolated cervical units. [31, 58, 59] Ultimately, the decision was to include only human data because the clinical application of the findings would be pertinent to the human population only. Computer-generated data and scaling articles were excluded to concentrate on primary human pediatric and adult data.
The data of tensile strength values of the spine were gathered from 9 articles, and the proposed ratios were based on 47 adult and 46 pediatric specimens that may be not be representative of the whole human population. However, 93 specimens could provide sufficient variability to be representative of the overall population. Several studies were excluded because of some reported data on the skull of pediatric patient and on frontal impact of the pediatric chest; a study on patients older than 65 years old was also excluded because the values differed from the adult population older than 65 years and because the data reported in adult studies were not comparable with the pediatric data.
Another limitation is that the ratios were based on the osteoligamentous spine and not on full spine that may be more representative of the living human patient. However, the 0.16 ratio obtained from intact spinal values (limitation from full spinal column in the pediatric case in comparison with full cervical spine) [27, 31] may be considered similar to the one obtained in the osteoligamentous spine (0.11). Therefore, it may be plausible that the ratios obtained in the osteoligamentous spine could be applicable in intact specimens.
The use of ratios is similar to the technique used by Hilker et al  to divide the adult population in small, mid, and large human sizes based on the percentile of their weights and heights. One of the limitations of presenting pediatric specimens as a scaled-down version of the adult ones is that the biomechanical differences may not be taken into consideration. Indeed, pediatric patients are not a scaled-down version of their adult counterparts but rather have a unique physiology that affects their biomechanical properties. [60, 61] However, the selection criteria of the review were not based on a scaling of pediatric specimens based on anthropometric data (height, weight) but specifically addressed the differences in biomechanical tensile properties in the selected specimens. Therefore, the proposed ratios were developed as an expression of the combined biomechanical and anthropometric differences observed between specimens.
There are limitations to the safety implications and the amount of “safe” loading described because they are relying on the findings of 2 studies reporting a direct cause-effect between vegetative responses and mechanical stimulation of the cervical spine. [14, 54] Vegetative responses may also be considered a normal physiological response to a mechanical stimulus in the cervical spine and may therefore not be representative of tissue damage. The reported episodes of bradycardia and apnea were arbitrarily considered the manifestation of adverse reactions secondary to treatment and not the occurrence of normal secondary consequences of the therapy implemented.  Indeed, the ratios would seem to indicate that 50 N of force might have been excessive for the population studied, and Koch et al  implied a link between vegetative responses secondary to mechanical stimulation and the initial stages of sudden infant death syndrome.
Limitations to the application of both safety and technique adaptation concepts and of the proposed model of pediatric SMT include the wide variability in the amount of forces used during manual therapy depending on the area treated and the clinician in adults. [2, 47] The few data on loads applied to the spine during SMT carry limitations on the adaptability of the concepts and model to the overall population of clinicians and patients. This preliminary model may evolve in the future as more relevant information is reported in the literature.
Currently, the amount of force used to treat children older than 12 weeks with pediatric SMT is unknown. Future research may use reliable measurement instrumentations or skin tensile pads to investigate forces used in pediatric SMT.  The validity of the proposed model of care remains to be proven, although the little evidence currently available seems to support it. Biomechanical studies are difficult to perform considering the rarity of specimens; however, investigating technique adaptations in various age groups (pediatric and adult) may also be appropriate to determine the validity of the proposed ratios and the model of care. Future research aiming at determining technique adaptation, apart from age, may want to consider other factors including weight, stature, and sex. Indeed, obesity has been shown to lead to advanced bone age, which was insufficient in overcoming greater forces generated by a larger mass, and overweight girls were found to have a lower bone mineral density compared with normal-weight girls. The overall risk of fracture was increased by 2 in children with higher fat mass.  Stature and body mass were found to have significant relationships with muscular strength (isokinetic strength).  Sex had an impact on reaching functional bone stiffness: boys increase in size, whereas girls increase in intrinsic material properties (volumetric bone mineral density). One study reported a difference in cervical axes of motion based on sex,  whereas another reported no sex differences in range of motion.  Considering that the tensile properties of the cervical spine are dependent on bone, ligaments, and muscle tissue, body mass index and sex may also be of importance when considering pediatric SMT technique adaptations. 
The information gathered from studies identified in the literature review suggests a nonlinear increase in the tensile strength of the cervical spine according to the age of human specimens. Based on the reported tensile strengths, a preliminary model of care combining the scaling ratios and the reported technique adaptations used during pediatric SMT has been proposed. The safety and clinical implications of the preliminary model of care may have an impact on the practice of SMT for infants and children.