Open questions — degenerative disc disease¶
Last curated: 2026-08-30
Stable identifiers are permanent. A question can be narrowed, answered or retired, but its OQ number should not be reassigned.
Tier 1 — would change practice and is designable now¶
OQ-1 — Can a noninvasive model identify a painful disc or endplate?¶
Why open. Degeneration is highly prevalent without symptoms, symptomatic association is probabilistic, and discography lacks an independent reference standard (Brinjikji 2015, PMID 25430861; Brinjikji 2015, PMID 26359154; Willems 2013, PMID 23427903).
Design now. Prospective multicenter cohort with locked clinical, conventional/quantitative MRI and biomarker model; independent adjudication; treatment-naïve prediction; external validation.
Practice change. Reduce incidental labeling and improve selection for targeted procedures.
OQ-2 — Are endplate-driven and annulus-driven degeneration reproducible clinical phenotypes?¶
Why open. A two-phenotype model explains divergent level, age and loading patterns but remains a mechanistic synthesis rather than a prospectively validated classifier (Adams 2012, PMID 22881295).
Design now. Repeated MRI/CT plus pain phenotyping from adolescence through adulthood; blinded phenotype assignment; predeclared transitions.
Practice change. Separate prevention and treatment pathways now grouped as DDD.
OQ-3 — Which Modic changes are causal, consequential or incidental?¶
Why open. Community data support type 2 change as a sequel of degeneration; systematic and phenotype-focused reviews disagree about pain association (Teichtahl 2017, PMID 27324605; Hopayian 2023, PMID 36438174; Czaplewski 2023, PMID 37170132).
Design now. Population cohort with lesion area/type transitions, standardized symptoms, endplate measures and blinded imaging core.
Practice change. Clarify diagnostic language, antibiotic avoidance and BVN-ablation eligibility.
OQ-4 — Does basivertebral nerve ablation retain a specific benefit in independent sham-controlled replication?¶
Why open. The pivotal sham trial showed ODI change −20.5 versus −15.2 and responders 75.6% versus 55.3%, while later evidence is sponsor-linked or lacks a persistent randomized sham control (Fischgrund 2018, PMID 29423885; Fischgrund 2020, PMID 32451777).
Design now. Publicly funded multicenter sham RCT with preregistered analysis, five-year registry linkage and trial-criteria transport cohort.
Practice change. Establish absolute specific benefit and coverage thresholds.
OQ-5 — Where should BVN ablation sit relative to optimized rehabilitation and fusion?¶
Why open. BVN trials used selected Modic 1/2 populations; fusion evidence uses different diagnostic constructs and rehabilitation comparators (Khalil 2019, PMID 31229663; Phillips 2013, PMID 23334400).
Design now. Pragmatic three-strategy trial or target-trial emulation in shared equipoise, with crossover and reintervention as outcomes.
Practice change. Define sequencing and avoid both premature surgery and serial low-value care.
OQ-6 — Which conservative program works best for a defined disc/endplate phenotype?¶
Why open. Exercise improves chronic nonspecific low-back pain on average, but trials rarely preserve DDD phenotype or demonstrate one superior exercise family (Hayden 2021, PMID 34580864; Fleckenstein 2022, PMID 35914641).
Design now. Factorial or adaptive RCT stratified by annular/endplate/radicular phenotype, with exercise dose and adherence captured.
Practice change. Replace generic referrals with evidence-based matching.
OQ-7 — Can imaging-report language reduce nocebo and low-value escalation?¶
Why open. Radiological labels affect beliefs and vulnerability, and both clinicians and patients overvalue imaging (Alhowimel 2022, PMID 32746671; Sharma 2020, PMID 32830105).
Design now. Cluster RCT of age-calibrated probabilistic reports plus clinician/patient decision support; measure fear, activity, repeat imaging, procedures and missed pathology.
Practice change. Turn reporting into a safe clinical intervention.
OQ-8 — Does lumbar disc replacement reduce symptomatic adjacent disease over a device lifetime?¶
Why open. Randomized studies show selected five-year outcome/reoperation advantages, but adjacent degeneration persists and eight-to-ten-year evidence does not resolve lifetime revision burden (Zigler 2012, PMID 23082846; Radcliff 2018, PMID 28005616; Furunes 2018, PMID 29794581).
Design now. Device-specific registry linkage with fusion and rehabilitation comparators, competing risks and ≥15-year follow-up.
Practice change. Inform implant choice for younger adults.
OQ-9 — Which patients gain more from fusion than intensive rehabilitation?¶
Why open. Mean surgical improvement is substantial, but randomized comparative advantage is modest/variable and prognostic level tests perform poorly (Chou 2009, PMID 19363455; Willems 2013, PMID 23427903; Koenders 2019, PMID 29995169).
Design now. Individual-participant meta-analysis followed by prospective validation of prespecified treatment-effect modifiers.
Practice change. Replace “failed conservative care + MRI” with a validated benefit model.
OQ-10 — Can allogeneic disc progenitor-cell efficacy be independently replicated?¶
Why open. A 60-person randomized trial showed dose-specific disc-volume and pain improvements, but the groups were small and the product has not been independently replicated (Gornet 2024, PMID 38925869).
Design now. Independent phase IIb/III sham-controlled trial with standardized rehabilitation, imaging core, product potency assay and ten-year safety follow-up.
Practice change. Determine whether a regulated disease-modifying product exists.
OQ-11 — Why did cell products produce conflicting randomized results?¶
Why open. RESPINE found 74% versus 69% primary responders (p=0.77), whereas other precursor/progenitor trials report signals; products, phenotype and endpoints differ (Pers 2024, PMID 39393844; Amirdelfan 2021, PMID 33045417; Beall 2025, PMID 40174800).
Design now. Harmonized individual-level consortium analysis with product characterization, endplate phenotype and negative-result inclusion.
Practice change. Prevent invalid “stem-cell class effect” claims.
OQ-12 — What paired structural and clinical endpoint defines regeneration?¶
Why open. Quantitative MRI is not a validated surrogate, and disc volume may change without native mechanics or clinical mediation (Russo 2023, PMID 37247638; Gornet 2024, PMID 38925869).
Design now. Prospective multi-trial surrogate validation with pain, ODI, activity, reintervention, composition and mechanics.
Practice change. Set regulatory and publication standards for regenerative claims.
OQ-13 — Does endplate transport predict response to intradiscal therapy?¶
Why open. Nutrient delivery is mechanistically central, but human DCE-MRI evidence includes a nine-patient study and no treatment interaction (Urban 2007, PMID 17260404; Muftuler 2015, PMID 25421547).
Design now. Standardize perfusion/transport imaging and prospectively test interaction in a randomized cell/hydrogel trial.
Practice change. Avoid injecting cells into biologically nonviable discs.
OQ-14 — Can PRP efficacy survive a standardized sham-controlled trial?¶
Why open. Small trials and reviews mix preparation, dose, discography and phenotype; results are inconsistent (Chang 2020, PMID 32607308; Akeda 2022, PMID 35053999; Zielinski 2022, PMID 39238810).
Design now. Multicenter sham RCT with central product assay, locked phenotype and two-year follow-up.
Practice change. Establish whether intradiscal PRP should remain research-only.
OQ-15 — Can biacuplasty be independently replicated against a credible sham?¶
Why open. Positive placebo-controlled and active-care trials are small and depend on the uncertain discogenic construct (Kapural 2013, PMID 23279658; Desai 2016, PMID 26689579).
Design now. Independent sham RCT with discography protocol sensitivity analysis and blinding assessment.
Practice change. Clarify a currently heterogeneous procedure pathway.
OQ-16 — Which adolescence trajectory predicts adult symptomatic disease?¶
Why open. A small 26-year cohort found the fastest Pfirrmann-score increase during ages 11–19 and more extensive age-19 disease among those with lifetime pain at 34 (Lund 2026, PMID 41786224).
Design now. Large school-based multiethnic cohort with growth, sport, metabolic, psychosocial and repeated MRI data.
Practice change. Identify or reject a prevention window.
OQ-17 — What is the causal contribution of smoking, obesity and occupational loading to each phenotype?¶
Why open. Twin studies show small exposure effects, while a large postmenopausal cohort found heavy loading associated with L5–S1 severe degeneration (OR 1.86, 95% CI 1.19–2.92) (Battié 2009, PMID 19111259; Salo 2022, PMID 35084078).
Design now. Wearable/measured exposure cohort with genotype, metabolic biomarkers, level-specific imaging and causal modeling.
Practice change. Improve prevention and workplace design without deterministic blame.
OQ-18 — Which red-flag combinations safely trigger emergency imaging?¶
Why open. Individual red flags are often insensitive and serious pathology prevalence differs sharply by setting (Galliker 2020, PMID 31278933; Dionne 2019, PMID 31132655).
Design now. Prospective international diagnostic cohort with standardized reference follow-up and decision-curve analysis.
Practice change. Reduce both missed emergencies and indiscriminate imaging.
OQ-19 — Can cauda equina syndrome definitions and pathways be standardized?¶
Why open. Incidence estimates vary from community to suspected cohorts, with heterogeneous definitions and predominantly UK single-center data (Hoeritzauer 2020, PMID 32059184).
Design now. International registry using consensus clinical/radiological stages, symptom onset, decompression timing and patient-reported outcomes.
Practice change. Make time-critical service design evidence-based.
OQ-20 — What patient-prioritized core outcome set should DDD trials use?¶
Why open. Low-back-pain core sets exist, but DDD trials span rehabilitation, devices, ablation and regeneration with incompatible structural and clinical endpoints (Chiarotto 2017, PMID 28460714; Bombardier 2001, PMID 11246692).
Design now. International Delphi with patients, clinicians, regulators and payers followed by feasibility testing.
Practice change. Reduce selective reporting and improve cross-trial comparison.
Decision-grade specifications for the Tier 1 portfolio¶
The table makes the principal estimand and a result that would close or materially redirect each question explicit. “Positive” is not defined as nominal statistical significance alone.
| OQ | Primary estimand / comparison | Minimum discriminating design feature | Result that would materially redirect the question |
|---|---|---|---|
| OQ-1 | External-validation calibration and decision-curve net benefit for painful-level prediction | Asymptomatic and symptomatic controls; locked model; no treatment response in the reference label | Calibration failure or no net benefit over age, examination and conventional MRI |
| OQ-3 | Within-person Modic transition → subsequent pain/function change | Serial lesion area/type; time-varying confounding; blinded symptom collection | No temporal dose–response after confounder control despite adequate transition count |
| OQ-4 | Sham-specific difference in ODI and pain responders at 12 and 24 months | Independent funding; credible blinding; treatment-policy analysis | Confidence interval excludes the prespecified smallest worthwhile effect |
| OQ-6 | Treatment-effect interaction between phenotype and conservative program | Factorial/adaptive allocation; dose and adherence measured | Replicated interaction absent, favoring phenotype-agnostic exercise choice |
| OQ-8 | Device-specific cumulative incidence of symptomatic adjacent disease and revision at ≥15 years | Competing-risk registry linkage; rehabilitation and fusion comparators | No clinically important motion-preservation advantage with equal or higher revision burden |
| OQ-9 | Individual absolute benefit of fusion over intensive rehabilitation | Treatment-effect model, not outcome prediction alone; external validation | Model fails calibration or identifies no reproducible subgroup with worthwhile incremental benefit |
| OQ-10 | Disc-progenitor product versus credible sham on dual clinical/structural endpoints | Potency assay; imaging core; prespecified mediation; long safety follow-up | Pain benefit without reproducible structure/function change, or vice versa |
| OQ-12 | Trial-level and individual-level causal surrogacy of quantitative MRI | Multiple treatment classes and at least one effective intervention | MRI change fails to mediate or predict patient-important benefit |
| OQ-18 | Decision-curve net benefit of a red-flag combination across care settings | Prospective consecutive recruitment and standardized reference follow-up | Rule increases missed serious pathology or imaging without net clinical benefit |
| OQ-20 | Completeness and feasibility of a patient-prioritized core set | Patient-led multilingual Delphi plus field testing | Persistent nonresponse or measurement burden makes the proposed set unusable |
Two new quantitative cautions alter the portfolio. ODI and RMDQ responder classification can vary markedly with derivation method—RMDQ improvement ranged from 14% to 51% in one 447-person study—so all responder-based questions must lock thresholds before analysis (Jordan 2006, PMID 16360560; Solomito 2025, PMID 38887023). Likewise, large high-intensity-zone odds ratios against discography do not solve OQ-1 because the reference construct and imaging feature are biologically coupled (Fang 2017, PMID 28178999).
Tier 2 — important, but prerequisite methods or evidence are missing¶
OQ-21 — Can disc-cell senescence be targeted safely in humans?¶
Senescence is mechanistically plausible, but human causal timing, delivery and off-target roles are unresolved (Feng 2016, PMID 27192096; Vo 2016, PMID 26890203). Prerequisites: validated in-vivo senescence marker and disc-targeted pharmacology.
OQ-22 — Which cell-death program dominates by disease stage?¶
Apoptosis, pyroptosis, ferroptosis, necroptosis and autophagy signatures overlap in preclinical work (Chen 2024, PMID 39639370). Prerequisite: replicated human spatial/single-cell atlas with longitudinal inference.
OQ-23 — Can a human disc cell atlas produce stable clinical endotypes?¶
Single-cell studies identify candidate fibrotic and inflammatory populations, but sampling and cluster labels are unstable (Ling 2025, PMID 40770348). Prerequisite: multi-center common-tissue protocol including non-surgical controls.
OQ-24 — Can gene or RNA therapy restore matrix without fibrosis or ectopic bone?¶
MicroRNA and noncoding-RNA targets are abundant but largely preclinical (Wang 2015, PMID 26368266; Ran 2022, PMID 35653923). Prerequisite: durable large-animal delivery and toxicology.
OQ-25 — Can annular closure make nucleus regeneration durable?¶
Hydrogels and cell products may extrude through fissures, while closure must survive cyclic torsion and tension (Zhang 2025, PMID 41084298; Ju 2020, PMID 32528797). Prerequisite: standardized mechanical failure tests.
OQ-26 — Does the disc circadian clock matter clinically?¶
Experimental evidence links circadian regulation to disc homeostasis, but no human chronobiology or treatment trial exists (Morris 2021, PMID 33397731). Prerequisite: time-standardized human biomarker and imaging studies.
OQ-27 — Can systemic biomarkers distinguish a painful level?¶
Inflammatory and matrix markers are nonspecific and cross-sectional (Khan 2017, PMID 29265416; Ding 2024, PMID 38528617). Prerequisite: local–systemic mapping and external validation.
OQ-28 — Can quantitative MRI become a regulatory surrogate?¶
Techniques detect composition but lack harmonization and causal surrogate validation (Russo 2023, PMID 37247638). Prerequisite: cross-scanner standards and multiple effective treatment classes.
OQ-29 — What determines persistent pain after structural treatment succeeds?¶
Fusion, replacement and biologic trials show discordance among technical, structural and clinical outcomes (Phillips 2013, PMID 23334400; Gornet 2024, PMID 38925869). Prerequisite: integrated local nociception, sensory processing, sleep and psychosocial phenotyping.
OQ-30 — How should mixed nociceptive, neuropathic and nociplastic features be represented?¶
DDD cohorts often mix axial and radicular disease while chronic pain processing modifies disability (Mohd Isa 2022, PMID 36613651; Ho 2022, PMID 35354560). Prerequisite: reliable multidomain classification with treatment interaction.
OQ-31 — What are the lifetime safety profiles of intradiscal cells and implants?¶
Two-to-five-year trials cannot exclude rare tumor, ectopic tissue, wear or revision risks (Gornet 2024, PMID 38925869; Zigler 2012, PMID 23082846). Prerequisite: mandated interoperable product registries.
OQ-32 — How can global equity be built into DDD research?¶
The global burden is largest and growing while primary country-level data are sparse; patient-voice literature is concentrated in high-income English-speaking settings (GBD 2021 Low Back Pain Collaborators 2023, PMID 37273833; Toye 2015, PMID 25793491). Prerequisite: multilingual community cohorts and locally feasible outcome measures.
Dots not yet connected¶
Each row identifies two established evidence bodies in this knowledge base for which a direct connecting study was not identified in the live searches through 2026-08-30.
| # | Dot A | Dot B | The missing junction | Powers |
|---|---|---|---|---|
| 1 | Accelerated disc change during puberty (PMID 41786224) | Adult genetic architecture (PMID 19111259) | Genotype-by-growth effects on level-specific trajectories | Longitudinal youth biobank + MRI |
| 2 | Endplate perfusion signal (PMID 25421547) | Cell-therapy response (PMID 38925869) | Does transport modify cell survival and benefit? | Biomarker-stratified RCT |
| 3 | Endplate-driven phenotype (PMID 22881295) | BVN-ablation efficacy (PMID 29423885) | Does mechanistic phenotype outperform Modic type for selection? | Prospective interaction analysis |
| 4 | Imaging-label nocebo (PMID 32746671) | Procedure utilization | Does report language causally change escalation? | Cluster RCT + claims linkage |
| 5 | Exercise benefit (PMID 34580864) | Disc quantitative MRI (PMID 37247638) | Do compositional changes mediate any clinical exercise response? | Mechanistic exercise RCT |
| 6 | Smoking/metabolic risk (PMID 1948392) | Endplate transport (PMID 25421547) | Is microvascular transport the human mediator? | Causal mediation cohort |
| 7 | Heavy occupational loading (PMID 35084078) | Two degeneration phenotypes (PMID 22881295) | Does measured load selectively drive annular rather than endplate disease? | Wearable exposure + serial MRI |
| 8 | Qualitative work burden (PMID 23585633) | Surgical/device outcomes (PMID 28005616) | Which operation effects translate into sustained, quality-adjusted work participation? | Registry + qualitative sub-study |
| 9 | Cell senescence (PMID 27192096) | Pain-source validity (PMID 23427903) | Is a senescent molecular state specifically nociceptive? | Spatial tissue–pain study |
| 10 | Quantitative MRI (PMID 37247638) | Discography uncertainty (PMID 18690280) | Can imaging replace invasive provocation without circular validation? | Blinded prospective diagnostic study |
| 11 | Medication small effects (PMID 37014979) | Patient medication stigma/experience (PMID 37696066) | Which shared-decision format improves net benefit and safe deprescribing? | Pragmatic communication RCT |
| 12 | Red-flag low sensitivity (PMID 31132655) | Imaging stewardship (PMID 25430861) | Decision rule optimizing missed pathology versus incidental cascade | International diagnostic-impact trial |
| 13 | Hydrogel mechanics (PMID 41084298) | Endplate fracture risk | Which material properties restore pressure without overloading endplates? | Human-scale organ + large-animal program |
| 14 | Disc replacement motion preservation (PMID 29794581) | Adjacent symptomatic disease | Does preserved motion mediate clinically meaningful adjacent outcomes? | Long-term causal mediation registry |
| 15 | Pain Alliance/organization priorities | Trial core outcomes (PMID 28460714) | Do DDD trials measure intimacy, work, access and treatment burden? | Patient-led core-set study |
| 16 | Polygenic propensity to multisite pain (PMID 38942297) | MRI degeneration–pain association | Does pain liability modify prospective symptom emergence rather than only cross-sectional association? | Genotyped serial-MRI cohort |
| 17 | Four human transcriptomic disc subtypes (PMID 40883814) | Conflicting cell-product RCTs (PMID 39393844) | Does molecular subtype modify cell-product survival or clinical effect? | Assay-stratified archived-sample RCT analysis |
| 18 | Diurnal disc deformation of 6–13% (PMID 35072794) | Quantitative MRI treatment endpoints (PMID 37247638) | How much apparent treatment change is acquisition-time variation? | Same-day repeatability substudy embedded in trials |
| 19 | Inappropriate imaging referrals (PMID 29730460) | Imaging-label nocebo (PMID 32746671) | Does reducing imaging also reduce fear and intervention cascades, or only cost? | Cluster policy trial with belief and utilization outcomes |
| 20 | Responder-rate method sensitivity (PMID 16360560) | Device/biologic composite success | Would treatment rankings change under common absolute, relative and acceptable-state thresholds? | Individual-participant reanalysis across product classes |