Working model

Hypotheses with what they fail to explain

Five working explanations, each shown with the findings it accounts for and the findings it does not. None is a diagnosis.

Analysis version v0.4.1 · evidence current through 2026-08-10 · Changelog

The models on the record

Ordered by stable identifier, not as a leaderboard. Jump to any full record.

H1Several separate problems stacked togethermedium+5 / −4
H2Could an inherited bone condition have been there from birth?medium+3 / −4
H3Calcium handling never measured under controlled conditions, and the bone-building markers are missingmedium+3 / −2
H4Could a lingering bacterial infection still be contributing?low+2 / −4
H5An unusual antibody patternlow+2 / −2
H-NULLNo single unifying diagnosisbaseline, not a rival

+ findings the model would explain · − findings it would not

How to read these records

Each record shows support and contradiction with equal visual weight, plus what would change the assessment. Counts are not weights, and the two columns are never netted against each other.

Confidence is five words only — never a percentage, score, or continuous bar. Dual architecture / module buckets are shown separately and never averaged.

not supportedspeculativelowmediumhigh

Rankings reflect multi-model research synthesis (Round 1, 2026-08-05), not a clinical diagnosis. Models can be wrong and can agree on errors.

Printable evidence table · How this could be wrong

H1Working hypothesisSource auditedas of 2026-08-05

Several separate problems stacked together

Layered skeletal–metabolic–neurologic–immune stack (infection detachable)

Preferred multi-disease research architecture from Round-1 synthesis — a research organizing frame, not a single causal mechanism and not a diagnosis. Infection limb detachable pending adjudication. Does not by itself explain tryptase/HαT (CLM-0030/0031), Babesia FISH×2 (CLM-0036), or historical profound hypogonadism (CLM-0018). Supporting cards are domain or methodological anchors (densitometry reporting, thiamine biomarkers, humoral abnormality), not proof of a layered multi-disease etiology. Models can agree on errors; architecture confidence is medium pending further adjudication.

This is a working research hypothesis, not a diagnosis. Rankings reflect multi-model research synthesis (Round 1, 2026-08-05), not a clinical diagnosis. Models can be wrong and can agree on errors.

+ What it would explain

CLM-0003: DXA lumbar L3 Z-score −4.3 on 2025-06-16 at Site 1 (BMD 0.631 g/cm²); same-day L3 T-score −4.3 (Z-score preferred for males <50 per ISCD).
CLM-0015: 24-hour urine calcium values 283 → 254 → 333 mg/24h (male; common threshold often 300 mg/day or 4 mg/kg — 2 of 3 collections below 300), with high urine sodium on latest panel (174 mmol/24h) and alkaline urine pH 6.53. Middle value (254 on 2025-09-30) was normal at the performing lab (ref 100–300) and high only against Litholink range 40–250; ranges differ by lab.
CLM-0023: Serum/plasma thiamine 7 nmol/L on 2022-09-29 (reference 8–30) — laboratory-low / deficient by that assay.
CLM-0027: IgG subclass 1 persistently low and IgG subclass 4 persistently high across multiple 2025–2026 draws (e.g., IgG1 324 with ref 382–929; IgG4 163.6 with ref 4–86); total IgG remains in range.
CLM-0013: L5–S1 bilateral pars defects with grade 1 spondylolisthesis and right L5 root impingement are documented on imaging summaries; SCS implanted March 2026.

− What it does not explain

CLM-0030: Serum tryptase elevated (17.7 µg/L on 2025-10-01; 14.5 mcg/L on 2026-06 panel) above ref <11; urinary mast-cell mediators normal on 2026 panel.
CLM-0031: Hereditary alpha-tryptasemia (HαT) reported positive in patient materials.
CLM-0036: Babesia FISH (whole blood) positive on two specialty-lab draws (Jul 2023 and Feb 2024); Babesia PCR (B. microti and B. duncani) and immunoblot IgM/IgG negative on Jul 2023 panel.
CLM-0018: Historical total testosterone reported as low as ~34 ng/dL (narrative/clinical-context in summaries); earliest discrete tabulated total T values in endocrine summary begin 2021.

Supporting literature (full list)

lit-0015: Executive Summary of the 2023 Adult Position Development Conference of the International Society for Clinical Densitometry: DXA Reporting, Follow-up BMD Testing and Trabecular Bone Score Application and Reporting (2023)
lit-0021: Thiamine deficiency disorders: diagnosis, prevalence, and a roadmap for global control programs (2018)
lit-0205: Specific Antibody Deficiency: Controversies in Diagnosis and Management (2017)

Counts are not weights. Supporting-list membership is not etiologic proof.

What would change this?

Concrete open items from the unresolved-question register. These are research gaps, not orders.

UQ-0003: Can original laboratory reports for total testosterone ≈34 ng/dL (reportedly ×2) be retrieved with dates, assay methods, and concurrent LH, FSH, prolactin, estradiol, SHBG?
UQ-0013: Has an early-onset osteoporosis / OI-spectrum gene panel been run, and with what result?
UQ-0017: Can specialty Babesia FISH / Bartonella IgM signals be independently adjudicated with reference-method testing (Babesia smear+PCR during symptoms; Bartonella culture/PCR/IFA/tissue if clinician-indicated), without relying on post-antibiotic FISH or IgM churn?

H2Working hypothesisSource auditedas of 2026-08-10

Could an inherited bone condition have been there from birth?

Monogenic / constitutional early-onset osteoporosis spectrum

Best single-primary research alternative (WNT1/LRP5/PLS3/mild OI spectrum; HPP low unless ALP truly low). Gene panel not performed. Related-cohort pathogenic yield is modest; panels can return variants of uncertain significance that complicate interpretation; negative panel does not close unknown genes or multi-hit secondary OP. Contradicting cards are now filed. lit-0340 covers what a bone-fragility gene panel actually finds in people referred for unexplained markedly low BMD for age (partial yield, so a genetic cause cannot be assumed from presentation), and lit-0341 covers male osteoporosis series in which identifiable secondary causes are common and more than one cause is often present at once. Both cut in the same direction as the caveats already named here — modest related-cohort yield and VUS interpretation difficulty — and neither can be read the other way, since a panel has not been done and so no result here argues for a genetic cause either. The 2026-08-10 Fable solo blinded round added two counterweights to this hypothesis's confidence, recorded here without changing its bucket: the record now carries an acquired exposure bearing on bone — an aromatase-inhibitor regimen recorded in the endocrine summary, a sensitive-assay estradiol of 9.3 pg/mL while on it, and about two years of weekly use per the patient's own account (CLM-0112; lit-0134, also filed on the contradicting list, records a year of that drug class lowering spine density in older men, while lit-0308 is carried as its short-horizon comparator — twelve weeks without marker worsening, which clears nothing about longer use) — and the patient's own account of decades of high-impact activity without peripheral fracture, which is an unusual biography for a monogenic fragility disorder of this severity, while also cautioning against reading the areal-BMD numbers as a complete description of bone strength.

+ What it would explain

CLM-0003: DXA lumbar L3 Z-score −4.3 on 2025-06-16 at Site 1 (BMD 0.631 g/cm²); same-day L3 T-score −4.3 (Z-score preferred for males <50 per ISCD).
CLM-0005: Left forearm total Z-score −3.1 and T-score −3.3 on 2026-06-19 Site 1 (BMD 0.516 g/cm²) (Z-score preferred for males <50 per ISCD).
CLM-0006: Total hip BMD fell from 0.836 to 0.802 g/cm² (compiled summary states −4.2%; recomputation (0.802−0.836)/0.836 ≈ −4.1%) between 2025-06-16 and 2026-06-19 on the same Site 1 scanner; compiled summary states LSC=0.027 g/cm² and calls the change significant at 95% confidence.

− What it does not explain

CLM-0023: Serum/plasma thiamine 7 nmol/L on 2022-09-29 (reference 8–30) — laboratory-low / deficient by that assay.
CLM-0027: IgG subclass 1 persistently low and IgG subclass 4 persistently high across multiple 2025–2026 draws (e.g., IgG1 324 with ref 382–929; IgG4 163.6 with ref 4–86); total IgG remains in range.
CLM-0036: Babesia FISH (whole blood) positive on two specialty-lab draws (Jul 2023 and Feb 2024); Babesia PCR (B. microti and B. duncani) and immunoblot IgM/IgG negative on Jul 2023 panel.
CLM-0112: The 2021 treatment record includes an aromatase inhibitor alongside the selective estrogen-receptor modulator: the endocrine summary documents the reported regimen during the 2021 titration period as clomiphene 25 mg daily plus anastrozole 1 mg weekly, and its hormone table prints an estradiol of 9.3 pg/mL on 2021-03-15 - measured by sensitive LC/MS/MS against an 8.0-35.0 pg/mL interval per its methodology note (5), while every other estradiol value in the table uses a standard immunoassay the same note says is not directly comparable. In the public video the patient states the aromatase inhibitor was taken "for two years" to lower a clomiphene-induced estradiol rise, and that both medications were then stopped without tapering before the 2021 symptom cascade (the automatic captions render the drug name as "an astrol"). This row records the exposure as history; it asserts nothing about what the exposure did.

Contradicting literature (full list)

lit-0340: Diagnostic yield of bone fragility gene panel sequencing in children and young adults referred for idiopathic primary osteoporosis at a single regional reference centre (2022)
lit-0341: Aetiology and clinical characteristics of male osteoporosis. Have they changed in the last few years? (2008)
lit-0134: Effects of aromatase inhibition on bone mineral density and bone turnover in older men with low testosterone levels (2009)

Supporting literature (full list)

lit-0287: WNT1 Mutations in Early-Onset Osteoporosis and Osteogenesis Imperfecta (2013)
lit-0290: PLS3 Mutations in X-Linked Osteoporosis with Fractures (2013)
lit-0294: Primary Osteoporosis in Young Adults: Genetic Basis and Identification of Novel Variants in Causal Genes (2018)Applicability flag — see card for cohort limits.
lit-0293: Mutational Analysis Uncovers Monogenic Bone Disorders in Women with Pregnancy-Associated Osteoporosis: Three Novel Mutations in LRP5, COL1A1, and COL1A2 (2018)
lit-0015: Executive Summary of the 2023 Adult Position Development Conference of the International Society for Clinical Densitometry: DXA Reporting, Follow-up BMD Testing and Trabecular Bone Score Application and Reporting (2023)

Counts are not weights. Supporting-list membership is not etiologic proof.

What would change this?

Concrete open items from the unresolved-question register. These are research gaps, not orders.

UQ-0013: Has an early-onset osteoporosis / OI-spectrum gene panel been run, and with what result?
UQ-0005: Do any historical ALP values fall below age- and sex-adjusted reference intervals (not only the adult 36–130 U/L band), and was bone-specific ALP ever measured?
UQ-0020: What did estradiol actually do across the roughly two years of weekly an aromatase inhibitor (per the patient's video statement) - are there primary laboratory reports with assay method for that window, and was the 2021-03-15 sensitive-assay value of 9.3 pg/mL (8.0-35.0) typical of the exposure or a single low point?

H3Working hypothesisSource auditedas of 2026-08-10

Calcium handling never measured under controlled conditions, and the bone-building markers are missing

Calcium-flux / incomplete bone-turnover phenotyping module

Ongoing calcium-flux research module (serial 24h urine Ca 283→254→333 mg/24h in a male; common male threshold often 300 mg/day or 4 mg/kg — 2 of 3 collections below 300; middle value 254 normal at performing lab ref 100–300 vs Litholink 40–250) plus high urine Na and missing formation markers; module not unifier. CTX alone upper-normal without paired P1NP/BSAP does not prove uncoupling. Cited IH literature (lit-0152/lit-0153) describes mild BMD reduction and rate-of-loss association in stone formers — it does not account for spine T −4.3 magnitude. On stone status the record says two things and both belong here. The CT urogram of 2026-02-27 found no stones, and the urology summary records one lifetime stone, passed spontaneously in his late twenties after about a week of heavy energy-drink and fast-food intake. So this is a stone-forming history at its mildest end, not an absence of one, and the stone-former literature applies weakly rather than not at all. Three contradicting cards are now filed (lit-0337, lit-0338, lit-0339) covering whether urinary calcium tracks bone-density severity at all and whether stone-related fracture excess is site-selective. The last of those counts fractures rather than measuring bone density, so it bears on this module only through a specific step, stated here rather than left to be inferred — the excess it reports is confined to the spine, while the low readings in this record appear at spine, hip and forearm at once, which is more than a stone-and-calcium mechanism has been shown to produce. Magnitude mismatch and incomplete mechanism classification remain the other named counters. Distal renal tubular acidosis — the textbook differential for alkaline urine with high urine calcium and metabolic bone disease, and until now never named anywhere in this record — is registered here as considered and not favoured: the same stone-risk panel shows urine citrate 784 mg/24h against a lower limit of 400 (CLM-0088) where that condition characteristically drives citrate low, and serum total CO2 sits within its reference interval on every documented draw (CLM-0110), where the classic form requires a metabolic acidosis. Neither value excludes an incomplete form with preserved citrate; whether the alkaline-urine picture warrants formal acidification testing is a clinician judgment this record refers rather than closes.

+ What it would explain

CLM-0015: 24-hour urine calcium values 283 → 254 → 333 mg/24h (male; common threshold often 300 mg/day or 4 mg/kg — 2 of 3 collections below 300), with high urine sodium on latest panel (174 mmol/24h) and alkaline urine pH 6.53. Middle value (254 on 2025-09-30) was normal at the performing lab (ref 100–300) and high only against Litholink range 40–250; ranges differ by lab.
CLM-0010: Serum CTX 616 pg/mL on 2026-06-26 (reference 70–780 pg/mL) — upper end of reference interval; fasting/time-of-day not stated in compiled summary.
CLM-0011: Bone formation markers (P1NP, bone-specific ALP, osteocalcin) are not documented among ~118 tests in the public ledger.

− What it does not explain

CLM-0023: Serum/plasma thiamine 7 nmol/L on 2022-09-29 (reference 8–30) — laboratory-low / deficient by that assay.
CLM-0027: IgG subclass 1 persistently low and IgG subclass 4 persistently high across multiple 2025–2026 draws (e.g., IgG1 324 with ref 382–929; IgG4 163.6 with ref 4–86); total IgG remains in range.

Contradicting literature (full list)

lit-0337: Relationship between Urinary Calcium and Bone Mineral Density in Patients with Calcium Nephrolithiasis (2017)
lit-0338: Determinants of osteopenia in male renal-stone-disease patients with idiopathic hypercalciuria (2011)
lit-0339: Fracture risk among patients with urolithiasis: a population-based cohort study (1998)

Supporting literature (full list)

lit-0013: Idiopathic hypercalciuria: Can we prevent stones and protect bones? (2018)
lit-0152: Urine calcium excretion predicts bone loss in idiopathic hypercalciuria (2006)
lit-0014: Rapid recovery of bone mass in hypercalciuric, osteoporotic men treated with hydrochlorothiazide (1999)
lit-0153: New insights into the pathogenesis of idiopathic hypercalciuria (2008)

Counts are not weights. Supporting-list membership is not etiologic proof.

What would change this?

Concrete open items from the unresolved-question register. These are research gaps, not orders.

UQ-0002: Was serum CTX 616 pg/mL (2026-06-26) drawn fasting, at a standardized morning time, and on which assay platform?
UQ-0005: Do any historical ALP values fall below age- and sex-adjusted reference intervals (not only the adult 36–130 U/L band), and was bone-specific ALP ever measured?

H4Working hypothesisSource auditedas of 2026-08-05

Could a lingering bacterial infection still be contributing?

Bartonella-spectrum residual contributor (specialty LDT contested)

Detachable residual contributor hypothesis for periosteal/nodule signals and contested specialty serology. Specialty IgM contested; PCR/FISH negative. Not commercially confirmed. Support is limited to applicability-limited musculoskeletal/osseous Bartonella case literature, not serology alone. Does not explain magnitude of generalized low BMD, historical hypogonadism, or hypercalciuria. Patient-reported improvement on atovaquone + clindamycin (± azithromycin) (CLM-0039) is a Babesia-directed regimen and is not organism-specific evidence for Bartonella — not listed as explained by H4.

+ What it would explain

CLM-0037: Bartonella immunoblot IgM genus/species positive Jul 2023, later indeterminate/negative pattern Feb 2024; Bartonella PCR and FISH negative.
CLM-0040: MRI left humerus Feb 2026: mild periosteal edema/enhancement without marrow signal change; reactive-appearing axillary nodes.

− What it does not explain

CLM-0003: DXA lumbar L3 Z-score −4.3 on 2025-06-16 at Site 1 (BMD 0.631 g/cm²); same-day L3 T-score −4.3 (Z-score preferred for males <50 per ISCD).
CLM-0018: Historical total testosterone reported as low as ~34 ng/dL (narrative/clinical-context in summaries); earliest discrete tabulated total T values in endocrine summary begin 2021.
CLM-0015: 24-hour urine calcium values 283 → 254 → 333 mg/24h (male; common threshold often 300 mg/day or 4 mg/kg — 2 of 3 collections below 300), with high urine sodium on latest panel (174 mmol/24h) and alkaline urine pH 6.53. Middle value (254 on 2025-09-30) was normal at the performing lab (ref 100–300) and high only against Litholink range 40–250; ranges differ by lab.
CLM-0039: Patient-reported joint pain, night sweats, nightmares, and REM sleep (~15 → ~90 min/night) improved during two courses of atovaquone + clindamycin (± azithromycin); the specialty summaries describe each course as producing sustained (≥6 month) improvement, with joint-pain relapse after the first course and the second (6-week, + azithromycin) course described as larger and more durable. Durability beyond the second course is not documented in the available summaries. This does not by itself prove Babesia, Bartonella, or other infection.

Contradicting literature (full list)

lit-0049: Characterization of human immunoglobulin (Ig) isotype and IgG subclass response to Bartonella henselae infection (1998)
lit-0042: Pitfalls and fallacies of cat scratch disease serology: evaluation of Bartonella henselae-based indirect fluorescence assay and enzyme-linked immunoassay (1997)
lit-0047: Limitations of Serological Diagnosis of Typical Cat Scratch Disease and Recommendations for the Diagnostic Procedure (2023)
lit-0055: Bartonella spp. seroprevalence in tick-exposed Swedish patients with persistent symptoms (2021)

Supporting literature (full list)

lit-0037: Musculoskeletal manifestations of cat scratch disease (2007)
lit-0050: Osteomyelitis caused by Bartonella henselae genotype I in an immunocompetent adult woman (2003)

Counts are not weights. Supporting-list membership is not etiologic proof.

What would change this?

Concrete open items from the unresolved-question register. These are research gaps, not orders.

UQ-0017: Can specialty Babesia FISH / Bartonella IgM signals be independently adjudicated with reference-method testing (Babesia smear+PCR during symptoms; Bartonella culture/PCR/IFA/tissue if clinician-indicated), without relying on post-antibiotic FISH or IgM churn?

What would not change this

Repeat standalone specialty IgM testing alone would not independently confirm infection and would not close commercial/PCR discordance.

H5Working hypothesisSource auditedas of 2026-08-05

An unusual antibody pattern

Humoral abnormality / possible specific antibody deficiency

IgG1-low / IgG4-high lab pattern is real and polyclonal. Named specific antibody deficiency remains low confidence until same-lab pre/post polysaccharide vaccine challenge and infection history. lit-0206 (Lawrence & Borish 2022) is a diagnostic-pitfalls review cautioning against overcalling SAD from titers alone — filed under contradicting literature, not support. Not classic IgG4-RD — lit-0089 and lit-0096 support that non-classic / elevation-without-RD framing (catalog polarity supports H5's position) rather than contradicting it.

+ What it would explain

CLM-0027: IgG subclass 1 persistently low and IgG subclass 4 persistently high across multiple 2025–2026 draws (e.g., IgG1 324 with ref 382–929; IgG4 163.6 with ref 4–86); total IgG remains in range.
CLM-0034: Hib IgG titer non-protective (0.33 mcg/mL, ref ≥1.00) and many pneumococcal serotype IgG values low on Jun 2026 panel; these are functional titers, not a formal vaccine-challenge study.

− What it does not explain

CLM-0003: DXA lumbar L3 Z-score −4.3 on 2025-06-16 at Site 1 (BMD 0.631 g/cm²); same-day L3 T-score −4.3 (Z-score preferred for males <50 per ISCD).
CLM-0023: Serum/plasma thiamine 7 nmol/L on 2022-09-29 (reference 8–30) — laboratory-low / deficient by that assay.

Contradicting literature (full list)

lit-0206: Specific antibody deficiency: pearls and pitfalls for diagnosis (2022)

Supporting literature (full list)

lit-0205: Specific Antibody Deficiency: Controversies in Diagnosis and Management (2017)
lit-0089: IgG4-related disease (2012)
lit-0096: The 2019 American College of Rheumatology/European League Against Rheumatism classification criteria for IgG4-related disease (2020)

Counts are not weights. Supporting-list membership is not etiologic proof.

What would change this?

Concrete open items from the unresolved-question register. These are research gaps, not orders.

UQ-0007: What is the vaccine history (products, dates) behind Hib and 23-valent pneumococcal IgG titers, and which assay platform/lab produced the June 2026 panel?

Null model / base-rate honesty

H-NULL — Multi-independent axes (null model / no single unifier)

Qualitative null / base-rate baseline — not a competing disease label. Several common or semi-independent processes may coexist without one rare unifying diagnosis. A defensible numeric population prior for this exact multi-system presentation is not available from the sources currently cited; false precision is avoided. This record exists for Bayesian discipline when weighing H1–H5, not as a diagnosis.

H-NULL exists for Bayesian discipline when weighing H1–H5. It is not a diagnosis and not a vote against them.

What this page does not say

No hypothesis here is a diagnosis. Infection limbs are detachable pending adjudication. Specialty laboratory-developed tests remain not independently confirmed.