Relationship between hypermobility and pain

Have been looking into the relationship between hypermobility on the one hand and pain, disability and health on the other. In other words: do people with hypermobility have more pain and worse health dan people without hypermobility?

It seems that there have been quite a few population-based studies, mostly in children. They show that hypermobility is more common in young people, females and in non-western countries. A 2012 review (McCluskey et al. 2012) found no association in European studies [OR 1.00, 95% confidence interval (CI) 0.79-1.26] but a marked relationship between hypermobility and joint pain in the Afro-Asian group (OR 2.01, 95% CI 1.45-2.77). I’m not sure what explains this major difference. One Indian study found malnutrition to be related to hypermobility so perhaps that complicates the relationship. Therefore, I decided to focus on studies in western countries for now.

Given that most people on the forum here are skeptical of a relationship, I think it might be worthwhile to focus first on the studies that reported a relationship. The McCluskey et al. 2012 review focused solely on cross-sectional studies but there are a couple of interesting longitudinal studies that reported hypermobility to be a predictor of future pain.

El-Metwally et al. 2004
One is a Finish study of 1756 schoolchildren with a mean age of 10.8 that were re-assessed 1 and 4 years later. Hypermobility was assed using Beighton scores. 7.8% had a score of 6 or higher. At baseline and 1-year follow-up there was no relationship between pain and hypermobility, but at the 4-year assessment there was. 76% of those with hypermobility had widespread musculoskeletal pain (mostly neck pain) compared to only 62.7% of those without hypermobility. In the multivariate analysis hypermobility had a risk ratio of 1.35 [1.08-1.68].
Prognosis of non-specific musculoskeletal pain in preadolescents: a prospective 4-year follow-up study till adolescence - PubMed (nih.gov)

The authors also used narrower pain definitions (e.g. lower limb pain) and here the results were more pronounced: 44% in hypermobile children versus 30% in controls had this pain resulting in an OR of 2.93 [1.13–7.70].
Lower limb pain in a preadolescent population: prognosis and risk factors for chronicity--a prospective 1- and 4-year follow-up study - PubMed (nih.gov)

Tobias et al. 2013
This is a UK study that looked at 2901 children aged 13.8 years that were re-examined when they were 17.8 years. 4.6% had Joint hypermobility defined as a Beighton score ≥6 at baseline. Joint hypermobility had an increased risk of moderately troublesome musculoskeletal pain at the shoulder (OR: 1.68 [1.04-2.72]), knee (OR: 1.83 [1.10,3.02]), and ankle/foot (OR: 1.82 [1.05, 3.16]) in models adjusted for sex, maternal education, and body mass index. Other sites such as the spine, elbows, hands, and hips showed no significant relationship. For chronic widespread pain overall, the OR was 1.84 [0.98, 3.46]. However, the above measures used the number of participants reporting moderate pain. When the authors looked at pain intensity scores and pain interfering with daily activities, they found no significant differences.
Joint hypermobility is a risk factor for musculoskeletal pain during adolescence: findings of a prospective cohort study - PubMed (nih.gov)

Sohrbeck-Nohr et al. 2014
This Danish study looked at 301 8-10 year olds which were re-examined at age 14. The main outcome was arthralgia but only 4% of participants had it. So while the relative risk was 3.00, the confidence intervals was wide, not reaching statistical significance. Using the RAOS questionnaire, pain was significantly worse in the hypermobility group if hypermobility was defined as Beighton ≥4 but not with a score of 5 or 6 as a threshold. There were also no significant differences for RAOS quality of life and symptoms.
Generalized joint hypermobility in childhood is a possible risk for the development of joint pain in adolescence: a cohort study - PubMed (nih.gov)

Interested in hearing what others think of these studies, especially @Jonathan Edwards if you have time.

Will post some studies in adults afterwards.

 

I agree. We are looking for causal factors for a problem affecting probably less than one percent of the population - widespread pain not reasonably explained by normal stresses and strains on tissues. A figure of 67% must be covering normally encountered pains. I also wonder if the study was blinded.

 

That makes more sense!

 

Yes, the relation between major hypermobility and joint injury/pain later in life is well recognised in the ballet profession. A lot of ballet dancers are suited to the work because of significant gene recurvatum and hip mobility, but it is understood that knee and hind foot injuries are more common where there is hypermobility. But these problems are very specific. This is not an issue of generalised pain or fatigue without injury.

 

The relation of hypermobility to malnutrition in India is very interesting. In the past I have seen a number of very thin women from the Indian subcontinent with widespread pain who look malnourished. The first thought was always osteomalacia, which is known to be prevalent in those with certain diets. These Indian women's hands are routinely very mobile and they even have a different finger posture at rest. The problem in interpretation was always that the women with pain were always from relatively uneducated lower social class groups that spoke no English and the history was always in exactly the same words, as if it was some sort of cultural code. A lot of rheumatologists were suspicious that a number of them were women abused by their spouses. But maybe malnutrition has a direct effect on ligament development that we did not appreciate.

 

There seems to be a problem a bit like the Long Covid studies that say that 74% of people still have three symptoms eight months after Covid. The reality of ill health doesn't reduce to numbers like that.

I went through several epidemiological studies, mostly from Scandinavia and on adolescents I think, given to me by an intelligent paediatrician I have lost contact with. The conclusion I came to was that there did seem to be some relation of hypermobility to spinal/neck pain (maybe the same study) but not that great.

Even risk factors of 2.0 are not that exciting in this context. The risk factor for having painful MCP joints in RA compared to controls is probably 90. What impressed me at the end of my searching was how weak and inconsistent any association seemed to be even if there was one.

 

For what it is worth here are some of the negative cross-sectional studies I found in western children. Unclear why most of the studies investigating this relationship were in (very young) children.

Mikkelsson et al. 1996 (Finland same study as El-Metwally et al. 2004)
7.8% had Beighton score ≥ 6. 29.9% with hypermobility had pain compared to 32.3% in controls. Hypermobility in one joint was not associated with pain in that body area. Disability did not correlate with the total Beighton score.
Joint hypermobility is not a contributing factor to musculoskeletal pain in pre-adolescents - PubMed (nih.gov)

Harreby et al. 1999 (Denmark)
14% hypermobility, 19.4% had LBP. “General hypermobility at any level […] was not correlated with LBP in any degree.”
Risk factors for low back pain in a cohort of 1389 Danish school children: an epidemiologic study - PubMed (nih.gov)

Qvindesland et al. 1999 (Iceland)
27.7% had Beighton≥4. 55% of hypermobile patiens reported pain versus 52% of controls.
Articular hypermobility in Icelandic 12-year-olds - PubMed (nih.gov)

Ruberto et al. 2004 (Italy)
34% had hypermobility using Beighton ≥ 5. “The presence of joint hypermobility does not affect the functional ability and the physical and psychosocial well being of otherwise healthy children”
Functional ability and physical and psychosocial well-being of hypermobile schoolchildren - PubMed (nih.gov)

Leone et al. 2009 (Italy)
22% of children with musculoskeletal pain versus 23% of controls had hypermobility defined as Beighton ≥5 (OR 1.057, 95% CI 0.7 to 1.4). “Hypermobile children did not experience functional limitations in daily activities, and they were slightly more active than non-hypermobile children”
Joint hypermobility and its relationship to musculoskeletal pain in schoolchildren: a cross-sectional study - PubMed (nih.gov)

Juul-Kristensen et al. 2009 (Denmark)
"There was no difference in daily level and duration of physical activity and in frequency of musculoskeletal pain and injuries between those with and without GJH.”
Motor competence and physical activity in 8-year-old school children with generalized joint hypermobility - PubMed (nih.gov)

Smits-Engelsman et al. 2011 (Netherlands)
"2.3% of children had symptoms of joint pain, and 9.1% complained of pain after exercise or sports. Importantly, this percentage was independent of the Beighton score"
Beighton score: a valid measure for generalized hypermobility in children - PubMed (nih.gov)

Aartun et al. 2016 (Denmark)
6.5% had Beighton score ≥ 5 and 4.25% ≥ 6. Neck and back pain occurred > 50%. Hypermobility had no significant association or predictor value for pain.
Validity of Commonly Used Clinical Tests to Diagnose and Screen for Spinal Pain in Adolescents: A School-Based Cohort Study in 1300 Danes Aged 11-15 Years - PubMed (nih.gov)

Morris et al. 2017 (Australia)
48% had Beighton≥4 and 18.6% Beighton≥6. In boys there was an association with hypermobility and more than 3 pain sites, but not in girls and only for Beighton≥6 threshold. No associations were found for number of pain areas lasting >3 months for either boys or girls at either threshold.
Hypermobility and Musculoskeletal Pain in Adolescents - PubMed (nih.gov)

There are also quite a few positive ones (finding a significant relationship between pain hypermobility) but most of these are from India, Turkey, Pakistan, Egypte etc. These often report much higher rates of hypermobility than studies in wersten countries so not sure how relevant these are.

 

I think studies in children are driven by the fact that paediatricians see a lot of unexplained pain and have taken the possible role of hypermobility seriously because they always think of a child's problems in the context of growth and development. Being paediatricians, who tend to be assiduous and systematic, for good reason, they have done some research. Adult rheumatologists have probably not bothered because they can sell (literally) hypermobility as a diagnosis and they wouldn't want a study to come out negative.

 

In adults, I couldn’t find that many studies.

Larsson et al. 1995
This Swedish study on high-technology industrial plant found that 26% of workers with spinal hypermobility experienced back pain compared to 14% in those without hypermobility. ‘Spinal hypermobility’ might not be the best term because it refers to workers being able to place their palms flat on the floor without bending their knees. They previously did a similar study in musicians and argue that “hypermobility of a joint is an asset if the joint is involved in repetitive motion and is a liability if the primary role of the joint is to provide support.”
Benefits and liabilities of hypermobility in the back pain disorders of industrial workers - PubMed (nih.gov)

Mulvey et al. 2013
Then there is this major UK study that had 12,853 participants because they used a shortened 5-item questionnaire to assess joint hypermobility (in other words, participants had to assess it themselves). 18.3% was hypermobile, 16.3% had chronic widespread pain. Hypermobile participants were significantly more likely to report pain than controls participants, but the difference was not large: 18.5% versus 15.8%. In the regression model, participants with hypermobility were 40% more likely to report chronic widespread pain of grade III/IV (RRR 1.4, 95% CI 1.1–1.7). However, the proportion of variance of the outcome (pain) explained by hypermobility in this first univariate model was only 0.65%. The authors conclude: “The relationship was relatively modest and may be explained by unmeasured confounding factors such as psychological distress.”
Modest association of joint hypermobility with disabling and limiting musculoskeletal pain: results from a large-scale general population-based survey - PubMed (nih.gov)

Juul-Kristensen et al. 2017
A similar questionnaire-based study was done in Denmark with 1006 participants. 30% of participants had hypermobility as assessed with the 5-PQ questionnaire. Odds ratios for shoulder (1.45 [1.09–1.92]), hands wrists (2.24 [1.67–3.01]) were significant but not for neck (1.34 [0.99–1.82]) and elbows (1.29 [0.93–1.80]). The EuroQol score was lower in hypermobile people (80 versus 85).
Generalised joint hypermobility and shoulder joint hypermobility, – risk of upper body musculoskeletal symptoms and reduced quality of life in the general population - PMC (nih.gov)

So I think these show that there might be a relationship but that it will likely be quite small.

 

This paper gives a simplified overview of the hypothesis of how joint pain could lead to chronic and widespread musculoskeletal pain.
Placing joint hypermobility in context: traits, disorders and syndromes - PubMed (nih.gov)

 

I'd be curious about the history of the Beighton Scale. How was it decided that the hypermobility of the joints in the scale were indicative of something meaningful?

 

Exams and swotting - 4 years later may be a key timeframe
My eldest daughter had shoulder and neck pain attributed to being hunched over revising for exams . No hypermobility
Perhaps a confounder ?

 

It wasn't. The Beighton score is simply a sum of all the simple signs of mobile joints. There was never any research to show that it meant anything of medical importance. Over the years people have suggested other scores but with equally little evidence for any significance.

Doctors are not good at working out why they design these scores. There is a thing called the GALS screen for joint problems - Gait, Arms, Legs, Spine. It is quite good in some ways but it was designed by four doctors. Some were interested in minimal signs of a problem - is this joint healthy or do I need to look at it in more detail? Others were interested in disability - is the person having trouble doing important activities? These are completely different purposes and mixing them together gives a routine that fails in some areas.

Most diagnostic criteria are just made up according to what doctors think might be a good idea - not based on any evidence of specificity. That has changed a bit with things like the Glasgow Coma Scale but not in all areas.

 

A lot of people are unhappy with the Beighton scale from what I can tell. This article gives a good overview of the main criticism:
https://pubmed.ncbi.nlm.nih.gov/33738549/

It was once used in the 1970s and then people chose to use because it was already used a lot.

 

This Dutch paper is interesting because the authors used a representative sample, split them up according to Beighton scores and then looked for other differences between the groups such as collagen biochemistry. They found that those with joint hypermobility and symptoms had more skin extensibility and higher values of hydroxyproline.

However, the study was small (n = 15) and in the discussion section they write:

https://pubmed.ncbi.nlm.nih.gov/12612280/

 

Interesting to see that. The main complaint seems to be that the Beighton Score may not reflect 'generalised hypermobility'. The trouble is that generalised hypermobility can mean whatever you want it to mean. The conclusion in the abstract seems unfortunate. it suggests that clinicians are able to discern 'generalised hypermobility' without having any specific measure at all. That to me is taking things in the wrong direction entirely.

I think it might be much more useful to study a single measure like elbow hyper extensibility or the thumb to forearm trick. Thumb to forearm has the advantage that it either does or doesn't and you cannot fudge an angle. Moreover, only a small minority of people can do it. Unless it is totally idiosyncratic it ought to correlate with whatever correlates with a widespread shift in ligamentous laxity - such as pain. If it doesn't then I think that would be a strong negative.

The other way to do it might be a multiple obligatory threshold score. If the assumption is that pain is associated with some consistent tendency to ligamentous laxity in many places then you don't really want to lump together a variety of different patterns of laxity, because they are likely to involve different genes and associations. You want to define a particular pattern, or perhaps several distinct patterns. SO you might have a system like.

1. The person must be able to put hands flat on the floor when standing with feet a foot apart.
and
2. Both elbows show 10° hyperextension
and
3. The person can do the thumb trick

That would be a tough set, but you are much more likely to get useful data out of a tough set than a vague set. If an association is convincing for a tough set then you can look at laxer definitions.

 

Why those 3 “tricks”? My son can do 2 and 3 but not 1. He scored 6 or 7 on the Beighton scale. There were additional tricks as per this photo & he can have his feet together but one pointing forward and the other backward, plus other contortions - but they were irrelevant in terms of scoring.

 

I am not suggesting those tricks are better than others but to make any scientific and statistical sense out of a score it should be an 'and' score not an and/or score. If someone cannot do trick 1 then you cannot build a theory around them that also fits people with trick 1.

I used to be able to do the shoulder trick too - I could touch my ear on the same side as the arm that went round behind my head and under my chin to do it. I also had 90 degree extension of pinkies (so scored 2). But I am just a random person and I suspect that at least half of random people can do one trick or another. And at certain stages of growth even more children can do these tricks.