One Spectrum, Two Labels

‍

Pradeep Chopra, MD, MHCM

Why the HSD–hEDS Divide May Not Hold Up

A critical look at the 2017 classification and the evidence behind the boundary

Tags: EDS · hEDS · HSD · connective tissue disorders · nosology

Ask two clinicians to examine the same hypermobile patient and you may get two different diagnoses. One calls it hypermobile Ehlers-Danlos syndrome. The other calls it hypermobility spectrum disorder. The patient has not changed. The skin has not changed. What changed is who did the examining. That, in a sentence, is the problem with the boundary the 2017 classification drew between these two conditions — and why it may be time to reconsider it.

A boundary built on consensus, not biology

The history of connective-tissue nosology is, in large part, a history of premature precision. The 1997 Villefranche criteria defined six EDS subtypes; the hypermobile type (then EDS type III) rested entirely on clinical features, with no known genetic defect. Two decades later, the 2017 International Classification expanded EDS to thirteen subtypes, kept hEDS as a clinically diagnosed category, and introduced a brand-new entity — hypermobility spectrum disorder — for patients with symptomatic generalized joint hypermobility who fell short of the newly tightened hEDS criteria.

The stated aim was more precise characterization: to separate a “more complex multisystemic disorder” (hEDS) from “less severe” or more localized complaints (HSD). On its face, this looks like refinement. In practice, it institutionalized a categorical boundary that has never been biologically validated. The line between the two is drawn by consensus criteria, not by any independent molecular or pathophysiologic marker.

The evidentiary hole at the center

The most basic requirement for splitting one disease into two is a distinguishing feature external to the diagnostic criteria themselves — a biomarker, a causative variant, a tissue finding, a reproducible physiologic measurement. For the HSD–hEDS split, no such discriminant exists.

Consider what the evidence actually shows across the domains that would matter:

●      Genetics. Alone among the 2017 EDS subtypes, hEDS has no identified causative gene — and neither does HSD. No variant differentiates the two. Whole-exome studies find overlapping, not divergent, profiles.

●      Ultrastructure. Collagen fibril abnormalities are described most consistently in the genetically defined subtypes. Reports in hEDS are heterogeneous, variably reproduced, and lack diagnostic specificity. No electron-microscopic signature reliably separates hEDS from HSD.

●      Proteomics and biomarkers. Tenascin-X deficiency points to classical-like EDS, not hEDS. Collagen cross-linking ratios, lysyl-hydroxylation patterns, and MMP activity have been reported in small or heterogeneous cohorts, but findings are inconsistent and unvalidated. No serum, urine, or tissue marker distinguishes the two.

●      Autonomic and neuroimmune features. Orthostatic intolerance, POTS, small-fiber neuropathy, chronic pain, and mast-cell activation–like presentations appear across both labels. Prevalence estimates vary with referral bias, not with diagnosis.

Every domain that could distinguish the two shows overlap, not separation.

The criteria contradict themselves

The 2017 hEDS definition requires three things at once: generalized joint hypermobility (criterion 1), a set of systemic features (criterion 2), and exclusion of alternatives (criterion 3). The boundary with HSD lives almost entirely inside criterion 2 — and that is where the criteria are weakest.

Criterion 2 has three feature clusters, of which a patient must satisfy at least two. Cluster A asks for five of twelve musculoskeletal and cutaneous features. Cluster B asks for a first-degree family history of hEDS. Cluster C asks for musculoskeletal complications or chronic pain beyond three months. Three problems follow directly:

●      Cluster A is subjective. Marfanoid habitus, striae, piezogenic papules, “soft velvety skin” — these are examiner-dependent with documented poor inter-rater reliability. Whether the same skin is judged “velvety” can decide the diagnosis.

●      Cluster B is circular. A family-history requirement is a proxy for heritability, but no relative can be confirmed to have hEDS by any objective test either. Patients with de novo presentation or unavailable family are penalized for reasons unrelated to their biology.

●      Cluster C conflates severity with identity. Chronic pain and musculoskeletal complications are near-universal in patients presenting for hypermobility evaluation. Using them to separate diagnoses confuses how sick a patient is with which disease they have — a classic taxonomic error.

Two labels, one set of rules

hEDS — 2017 criteria

HSD — 2017 criteria

●      Generalized joint hypermobility (Beighton)

●      Criterion 2: at least 2 of 3 feature clusters (A, B, C)

●      Exclusion of alternative diagnoses

●      Subjective skin / musculoskeletal features

●      No genetic confirmation required

●      No biomarker required

●      Generalized joint hypermobility (Beighton)

●      Fails to meet hEDS criterion 2 fully

●      Exclusion of alternative diagnoses

●      Symptomatic presentation required

●      No genetic confirmation required

●      No biomarker required

Set side by side, the only substantive difference is the criterion 2 threshold. Meet it, and the diagnosis is hEDS; fall one feature short, and it is HSD. Both exclude the same conditions, require the same Beighton score, and require the same symptomatic presentation. The gap between them is a score on a subjective checklist, not a biological boundary.

The populations are indistinguishable

If the two diagnoses named different diseases, their populations should differ on something that matters. They do not. Demmler and colleagues (2019) compared pain, fatigue, functional limitation, and quality of life across hEDS and HSD cohorts from a national registry and found no significant difference on any primary or secondary outcome. Tofts and colleagues (2023) found equivalent symptom burden, dysautonomia prevalence, and psychological comorbidity between the groups.

The common defense — that hEDS is simply the “severe” end and HSD the “mild” end — is not supported by the comparative data, and it is self-defeating besides. A severity gradient within a spectrum is exactly what a single disorder produces. It is not evidence of a categorical divide.

Weighing the claims for separation

Distinguishing claim

Supporting evidence

Assessment

hEDS is more severe than HSD

No adequately powered comparative study shows a consistent severity differential

Not supported

hEDS has distinct genetic architecture

No causative gene for either; no discriminating variant profile found

Not supported

hEDS has more systemic involvement

Dysautonomia, chronic pain, and MCAS occur at equivalent rates in comparative cohorts

Not supported

HSD may progress to hEDS over time

Longitudinal reclassification documented in case series — implies the same underlying condition

Supports unification

Different treatment responses justify separation

No RCT or high-quality cohort shows differential therapeutic response

Not supported

Patients migrate across the line

The most telling evidence against the split is reclassification. Case series and clinic audits show that many patients initially labeled HSD are later relabeled hEDS — not because their condition changed, but because a different examiner applied subjective criteria differently, elicited more features on fuller history-taking, or the patient learned which features to report.

That is not how two distinct diseases behave. It is how one disease with poorly drawn boundaries behaves. In no other field would we call two conditions categorically distinct if patients routinely move between them based on examiner variation rather than any change in disease.

When patients cross a diagnostic line because the examiner changed, not the patient, the line is administrative — not biological.

The costs are not neutral

This boundary does real harm. In practice, hEDS reads as the “legitimate” diagnosis and HSD as the lesser one. Non-specialists often treat HSD as milder and less deserving of aggressive management, so patients with HSD get reduced access to multidisciplinary pain care, physiotherapy, autonomic workup, and specialist referral — despite equivalent burden. The effect is worst in pediatric and socioeconomically disadvantaged populations, where a first-contact clinician’s reading of the HSD label can quietly close the door to hEDS-level care.

The split also fragments research. Studies recruiting “hEDS patients” and those recruiting “HSD patients” are, in all likelihood, studying the same disorder at different depths of ascertainment. The apparent discrepancies this produces — in prevalence, comorbidity, treatment response — are artifacts of classification. It also starves gene-discovery efforts of the pooled sample sizes GWAS power requires.

And there is direct patient harm: being told you have “almost” hEDS, or receiving HSD when you were led to expect hEDS, is a documented source of distress, doctor-shopping, and eroded trust. A unified framework would not resolve the etiology, but it would remove an iatrogenic hierarchy that has no scientific basis.

The case for one spectrum

The 2017 developers were candid: the hEDS criteria were proposed on expert opinion in the absence of confirmatory data, and they explicitly called for research to validate or revise the boundaries. That caveat has been under-emphasized ever since, as the HSD–hEDS distinction hardened into something treated as established rather than provisional.

The right response to a boundary that cannot be validated, that produces identical populations, that suffers poor inter-rater reliability, and that causes measurable harm is not to preserve it while we wait. The burden of proof lies with those asserting two conditions are distinct. Absent a validated discriminant, the null hypothesis is that HSD and hEDS are the same disorder.

What the field needs is a single category — a heritable connective-tissue hypermobility disorder, or similar — spanning the full range from isolated symptomatic hypermobility to complex multisystem disease, with severity specifiers replacing the artificial categorical line. That would match the observed continuum, unify rather than fragment the research base, dissolve the harmful hierarchy, and honestly represent what we know: a coherent phenotypic spectrum of unknown, probably heterogeneous, molecular origin. Categorical subdivision can wait for the day we actually find markers to justify it.

‍

‍

References

1. Beighton P, De Paepe A, Steinmann B, et al. Ehlers-Danlos syndromes: revised nosology. Villefranche, 1997. Am J Med Genet. 1998;77(1):31-37.

2. Malfait F, Francomano C, Byers P, et al. The 2017 international classification of the Ehlers-Danlos syndromes. Am J Med Genet C Semin Med Genet. 2017;175(1):8-26.

3. Castori M, Tinkle B, Levy H, et al. A framework for the classification of joint hypermobility and related conditions. Am J Med Genet C Semin Med Genet. 2017;175(1):148-157.

4. Pope FM, Nicholls AC, Narcisi P, et al. Type III collagen mutations in Ehlers-Danlos syndrome type IV and other related disorders. Clin Exp Dermatol. 1988;13(5):285-302.

5. Byers PH. Collagens: building blocks at the end of the development line. Clin Genet. 2000;58(4):270-279.

6. Hakim A, Grahame R. Joint hypermobility. Best Pract Res Clin Rheumatol. 2003;17(6):989-1004.

7. Castori M, Tinkle B, Levy H, Grahame R, Malfait F, Hakim A. A framework for the classification of joint hypermobility and related conditions. Am J Med Genet C Semin Med Genet. 2017;175(1):148-157.

8. Schalkwijk J, Zweers MC, Steijlen PM, et al. A recessive form of the Ehlers-Danlos syndrome caused by tenascin-X deficiency. N Engl J Med. 2001;345(16):1167-1175.

9. Bowen JM, Sobey GJ, Burrows NP, et al. Ehlers-Danlos syndrome, classical type. Am J Med Genet C Semin Med Genet. 2017;175(1):27-39.

10. Ritelli M, Colombi M. Molecular genetics and pathogenesis of Ehlers-Danlos syndrome and related connective tissue disorders. Genes (Basel). 2020;11(5):547.

11. Tinkle B, Castori M, Berglund B, et al. Hypermobile Ehlers-Danlos syndrome: clinical description and natural history. Am J Med Genet C Semin Med Genet. 2017;175(1):48-69.

12. Gazit Y, Nahir AM, Grahame R, Jacob G. Dysautonomia in the joint hypermobility syndrome. Am J Med. 2003;115(1):33-40.

13. Cazzato D, Castori M, Lombardi R, et al. Small fiber neuropathy is a common feature of Ehlers-Danlos syndromes. Neurology. 2016;87(2):155-159.

14. Molderings GJ, Brettner S, Homann J, Afrin LB. Mast cell activation disease: a concise practical guide for diagnostic workup and therapeutic options. J Hematol Oncol. 2011;4:10.

15. Demmler JC, Atkinson MD, Reinhold EJ, Choy E, Lyons RA, Brophy ST. Diagnosed prevalence of Ehlers-Danlos syndrome and hypermobility spectrum disorder in Wales, UK. BMJ Open. 2019;9(11):e031365.

16. Tofts LJ, Simmonds J, Schwartz SB, et al. Pediatric joint hypermobility: a diagnostic framework and narrative review. Orphanet J Rare Dis. 2023;18(1):104.

17. Palmer S, Bailey S, Barker L, Barney L, Elliott A. The effectiveness of therapeutic exercise for joint hypermobility syndrome: a systematic review. Physiotherapy. 2014;100(3):220-227.

18. Gensemer C, Burks R, Kautz S, Judge DP, Lavallee M, Norris RA. Hypermobile Ehlers-Danlos syndromes: complex phenotypes, challenging diagnoses, and poorly understood causes. Dev Dyn. 2021;250(3):318-344.

‍