Generalized hypermobility is said to be present when many joints in the body show an increased range of motion [1]. Most children with hypermobility have no complaints and are referred to as children with asymptomatic generalized hypermobility [2]. Some children even benefit from excessive range of motion in their joints to succeed in gymnastic or acrobatic skills. Middleditch [3] however describes the difference between ballet dancers and gymnasts with ‘normal tissues’ and increased range of motion due to training and stretching versus hypermobile individuals with generally weaker tissue. This tissue is less resilient; joints are less stable and vulnerable to subluxation. Hypermobile ballet dancers have a higher incidence of injuries than non-hypermobile dancers [4]. The prevalence of asymptomatic generalized hypermobility in children has been variably and widely reported, between 3% and 30% [5–7]. Hypermobility is generally measured with the Beighton test [2, 8]. Importantly, there seems a strong association between disorders of hypermobility and Developmental Coordination Disorder (DCD) [9].

However, since in these both studies the initial entry was hypermobility, none of the studies mentioned neither yielded significant evidence for co-occurrence of hypermobility and poor motor performance in children referred for motor problems or for children in a general nonclinical population. Kirby and Davies [14] did look at the co-occurrence in the other direction. They compared children diagnosed with DCD (n = 27) and a typically developing group (TDC; n = 27) using a questionnaire evaluating the clinical signs of children with hypermobility. Thirty-seven per cent of the children with DCD had symptoms of hypermobility (JHS), compared with 7,4% in the TDC group. Signs such as bending the thumb back to touch the forearm, contorting the body into strange postures or being double jointed were considerably more prevalent in the group of children with DCD compared to their matched peers.

Another approach used is looking into the physical fitness of children with high and low motor competence [15]. One of the items used to assess physical fitness and flexibility of the hip joints, lower back and length of Hamstrings, was the sit and reach test. The group with low motor competence had poorer results in flexibility. This finding is supported by Hands et al. (2006) in children with motor learning difficulties, while Schott et al. did not find a relationship with flexibility, measured with the sit and reach test, in a group of children with DCD [16, 17].

It can be concluded that there is evidence (in retrospective or questionnaire studies) of a higher incidence of motor delay in children who are referred for hypermobility, even in the absence of an identified neurological deficit. There is also evidence of enhanced flexibility in children with DCD [12, 18, 19]. Since humans have a complex body with more joints and with that more degrees of freedom than needed to perform any particular task, a flexible and adaptable motor system is needed to control all joints in a given task [20]. A limited capacity to control the degrees of freedom, as seen in children with DCD, shows up as inefficient timing of antagonist activation resulting in prolonged movement time and a need for more corrective movements during target directed motion [21]. This may be one of the underlying causes of the development of poor postural control and movement skills [9]. Since hypermobility results in larger degrees of freedom of motion in joints, it is important to better understand the consequences of hypermobility and its related symptoms during the motor development of both typically developing children and those with DCD.

The present study was designed to elucidate the association between joint mobility and motor performance. To estimate range of motion (ROM) goniometry was used to objectively measure joint mobility. The first aim of the study was to verify whether children referred for DCD would present more often with hypermobility (as measured by the Beighton score) than children in a random sample [2, 8]. A second aim was to evaluate the relationship between motor performance (as measured by the MABC) and joint mobility in a random sample of children aged 3–16 years old, without medical or neurological complaint [13, 22]. Finally we tested whether the relation between mobility and motor performance was different between the children with poor motor coordination and the general-population sample. Based on the literature, it was hypothesized that hypermobility would be associated with lower motor performance in both groups, but to different degrees, with a stronger association in the DCD group.

The aim of this study was to determine the prevalence of hypermobility in children with Developmental Coordination Disorder and in a randomly selected group of typically developing children in a wide age range (3–16 years of age). It was investigated whether hypermobility and ROM were associated with motor performance in both groups and whether the relation was different between the DCD group in contrast to an age matched group. For this purpose a sample of 36 children referred for DCD to physical therapy intervention and a random sample of 352 typically developing Dutch children was examined. They were assessed with the MABC to establish their motor performance and for joint mobility with goniometry to establish an objective value for the Beighton score.

It is of interest that the prevalence of hypermobility was twice as high in children with DCD (64%) as compared with the prevalence of hypermobility in both the random and matched TD groups, which was also rather high (33%). These percentages were obtained using the cut-off points for hypermobility recommended by van der Giessen [25], which were more strict than those reported by Beighton [2, 8]. Nevertheless, the high prevalence of hypermobility found in our study of typically developing children is difficult to accept and is much higher than the one reported by Jessee et al., who reported 5-7% of school children to be hypermobile [32]. In their study of Swedish school children, Jansson et al. stated that if hypermobility is considered a variation of general joint laxity such a high prevalence of non-normality is difficult to accept [6]. Our data substantiate that statement. The findings of this study also corroborate with the recommendation of Smits-Engelsman et al. to choose the 7 on the 9 point Beighton scale as the cut-off in children over 6 years of age, when mobility is tested passively by goniometry [24]. This raises the question whether the prevalence in both groups is more acceptable with a cut-off of 7. If a 7-point cut-off is applied to the data in the present study, the prevalence of hypermobility would become 6% in the general population, which is regarded as an acceptable percentage. For the DCD-group the prevalence would be 28% if this higher cut-off were used. A posthoc analysis shows that with the new criterion within the DCD group the subgroups with (n = 10) and without (n = 26) hypermobility clearly differ on the MABC2 (mean score 2.8 (SD = 1.8) versus 4.4 (SD = 1.7), t(2,34) = 2.5, p = .019). So far there remains uncertainty in the literature about cut-off points and ways to measure hypermobility. We therefore strongly recommend an international agreement on cut-off points and the use of the standardized measurement of Beighton mobility manoeuvres, since the prevalence of hypermobility in both random and clinical populations is otherwise difficult to compare.

Regardless of the chosen cut-off scores the prevalence of hypermobility, at 28%, was proportionally still much higher in the group of children referred for DCD (χ ² = 17.93, p = .001). Based on the literature, it was anticipated that asymptomatic joint hypermobility would be associated with poor motor performance. Interestingly, this was not the case in the randomly selected group (r_p = −.07). Given results in the random group, there seems no association between joint mobility and motor performance among typically developing children. However, the negative correlation within the DCD-group was significant and moderately poor (r_p = −0.38, p = .02). These findings demonstrate that when poor motor performance occurs, especially in tasks dealing with control of the whole body, 14% of the performance can be explained by asymptomatic joint hypermobility (R ²  = 0.14). Our findings provide evidence that testing for hypermobility, preferably by a standardized Beighton protocol using a goniometer, should be part of the assessment for children referred for motor coordination problems, since this may include a variable to consider in setting intervention goals on controlling and strengthening or stabilising movements.

It is well known that dealing with more degrees of freedom augments the complexity of motor control processes [20, 21]. Moreover it could enlarge loads on postural control due to decreased joint stability. In particular, the increased knee extension should be taken into account from a postural control perspective. According to our findings this extension was significantly related to the MABC2 score (left knee r_s = −0.48, p < .01; right knee r_s = −0.37, p = .03).

Looking at children with DCD one wonders by what mechanism more mobility or larger degrees of freedom in a joint might co-occur with motor coordination problems. According to Maillard and Murray [11] the aspect of reduced proprioception from the joints in children with hypermobility might lead to poor control of joint movement and instability. According to Geuze [33] the major characteristics of poor control in DCD are inconsistent timing of muscle activation sequences, co-contraction and lack of automatization and slowness of response. These characteristics will make it more difficult to control hypermobile joints, since a lack of co-contraction and slowness of response will result in decreased and less well timed stability of the loaded joints. This current study supports the notion that having to deal with larger degrees of freedom in joints can co-occur with motor problems in children with DCD. During development children might find a way to strengthen and control their hypermobile joints. For hypermobile knees this idea is supported by the study of Greenwood et al. [34]. Electromyography showed significantly higher semitendinosus activation overall, and significantly higher co-contraction of Rectus femoris and Semitendinosus during less challenging tasks (two-leg standing) of hypermobile adult participants with JHS compared to a control group. In contrast it is also known that children with high motor proficiency may excel if they have mobile joints, likely because they can exploit the larger degrees of freedom by intensive training (e.g. gymnasts). It is obvious that the stability of a joint is not only dependent upon intact ligament structures but also on neuromuscular control and muscle tone [1].

The finding that the number of children that could get their hands on the floor with straight legs (χ ² = 1.86, p = .17) did not differ between the matched TD-group and DCD-group is of interest. This links to the findings of Cantell and Hands that increased flexibility is not found in children with poor motor performance while doing the sit and reach test [15, 16]. The most probable explanation is that both the Beighton component of standing with straight legs and touching the ground with both hands flat on the floor, and the sit and reach test are muscle length tests of the Hamstrings and not range of motion tests of a joint. It may be worthwhile to evaluate the validity of this item of the Beighton as to whether it gives appropriate information on hypermobility.

The combination of generating appropriate levels of force and having to deal with more degrees of freedom in a joint may negatively affect motor performance in hypermobile children with DCD. Having to generate increased force in antagonistic muscles and more co-contraction may also cause extra recruitment noise and therefore cause additional movement variability, which is frequently reported in movement patterns of children with DCD [35, 36]. This statement is supported by the findings of Smits-Engelsman et al. in which children with DCD could produce the same level of maximum finger force as typically developing children but have poor control over maintaining steady force levels as required in the force control tasks [37]. Such a lack of fine tuned force control will lead to larger errors in precise movements.

We realize the number of girls and boys was not totally matched in the matched TD group compared to the DCD group. Jansson et al. found in Sweden a significantly higher degree of general joint laxity in girls of all ages compared to boys (p < 0.05) [6]. In contrast, three other studies found no significant differences by gender. First Rikken-Bultman found no significant difference in the group of a primary school [5]. Secondly van der Giessen found no differences in the prevalence of both hypermobility and the connective tissue signs between boys and girls with p values ranging from 0.115 to 1.000 [25]. Thirdly Smits-Engelsman recorded no significant differences (p = .22) for sex and if analyzed per item, only hands on the floor (item 5) was different (t (1549) = 4.66, p < .001); with girls being more flexible than boys [24]. So we had no reason to analyze sex differences and it is unlikely that sex differences explain any of the group differences at this age (7–10 years old).

The number of tests in the analysis of the association between MABC and different ROM has to be considered. On the other hand, one of the two knee extension measures would survive a correction for multiple tests. The fact that the other is also significant at the 0.05 level is clearly not irrelevant. It therefore seems reasonable to just describe the association as ‘significant’.

It should be taken into account that the clinical sample is small and larger samples would have been favoured. Also it would be an advantage to evaluate motor performance in more depth than is possible with a standardized 8 item test, like the MABC. Obviously a disadvantage of cross-sectional and correlation studies is that it is impossible to infer the developmental aetiology of the association. Do children with poor motor coordination become more mobile because they lack stabilization of their joints and as a result they increase their mobility? (For instance through repetitive joint sprains or as a strategy to stabilize the joint). It is important to investigate whether these children may adopt an altered posture whereby they “rest” or “hang” on the hip capsule and hip ligaments rather than activating their Gluteus medius, which would f.e. cause pelvic obliquity and instability [3]. The same can be expected for the knee and hanging in knee ligaments rather than activating the Quadriceps, Hamstrings and Gastrocnemius. Is deficient coordination the eliciting factor for instance through delayed or inadequate sensori-motor loops? Or are there other mediating factors that induce the co-occurrence of hypermobility and DCD? Further prospective and intervention research should elucidate the possible relations between underlying factors.

Hypermobility, scored according to the cut-off scores of the Beighton protocol, is a frequent phenomenon. It is even more frequent in children with DCD. Although in the general population motor performance and joint mobility as measured ROM by goniometry are not related, this is the case in children referred for DCD. Hypermobility of the knee extension seem to co-occur with a lower percentile of the MABC. We argue that coping with hypermobility or larger degrees of freedom may be a disadvantage when motor coordination is deficient.

Clinicians need to consider joint mobility, measured according to the Beighton protocol, as a factor in their assessment of children referred for coordination problems to find out whether hypermobility should be considered an implicating factor in children with poor motor control.