8-OHdG & mtDNA in Aging / MCI / AD / PD — 2026-06-20 切換：本地版 ↗
01

Overview & Biomarker Chemistry 總論 & 化學特性

8-Hydroxy-2′-deoxyguanosine (8-OHdG) is the canonical biomarker of oxidative DNA damage: reactive oxygen species (ROS) attack the C8 position of guanine in DNA, generating 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxo-dG); base-excision repair excises it as 8-OHdG, which is subsequently excreted in urine, or circulates as free nucleoside in plasma/serum and cerebrospinal fluid (CSF). Its RNA counterpart 8-OHG (8-hydroxyguanosine) derives from ribonucleoside oxidation, and the two are structurally similar but biologically distinct — RNA oxidation impairs translation; DNA oxidation drives mutagenesis.

Measurement hierarchy: HPLC-ECD (gold standard, urine) > UPLC-MS/MS (CSF — highest sensitivity, distinguishes 8-OHdG from 8-OHG) [1] > ELISA (serum/plasma — widely used but cross-reactivity risk inflates estimates). Studies must explicitly state which analyte is measured; many early reports labelled both as "8-OHdG". [28]

重點 尿液 8-OHdG（HPLC/ELISA）為最常用的氧化 DNA 損傷非侵入性標誌物；CSF 8-OHdG（UPLC-MS/MS）特異性最高，直接反映中樞神經系統氧化狀態。8-OHG 為 RNA 氧化指標，不應與 8-OHdG 混用。
Drive 20Local PDFs identified
PubMed 124After 3 MeSH queries
OE 7OpenEvidence citations
Included 30After quality filter (Q1/Q2, PMID verified)
結論 This review covers 30 PMID-verified studies (2 meta-analyses/SR, 1 cohort meta-analysis, 19 original studies, 4 reviews) published 2002–2026. Time window: ≥ 2016 (landmark classics retained). OE audit: CAUTION (7/7 citations exist; Abe 2002 correctly relabelled 8-OHG ≠ 8-OHdG).
02

8-OHdG in Alzheimer's Disease & MCI AD / MCI 之 8-OHdG

證據顯示 CSF 8-OHG (RNA oxidation) is ~5-fold elevated in AD versus controls, with complete group separation in the landmark study (all 18 AD >250 pM; all controls <140 pM). [1] CSF 8-OHG positively correlates with MMSE (rs=0.67, P<0.01) and inversely with disease duration (rs=−0.48, P<0.05), suggesting peak oxidative RNA damage occurs in early-to-intermediate stages before extensive neuronal loss. Serum 8-OHG was not significantly altered — CSF levels reflect brain tissue rather than systemic oxidative state.

Serum 8-OHdG (DNA oxidation) shows a dose-response gradient Control < MCI < AD (P<0.05; n=100/121/131) with negative correlations with both MoCA and MMSE; serum 8-OHdG and SAA are positively correlated, suggesting co-activation of oxidative and inflammatory pathways. [2] Plasma 8-OHdG is elevated in early AD compared to controls, [5] and is independently associated with motoric cognitive risk (MCR) in a large Chinese aging cohort (OR 1.007 per unit, 95% CI computed from P=0.003; n=1,312). [6]

Urinary 8-OHdG is elevated in AD patients, [3] and correlates inversely with plasma paraoxonase-1 (PON1) activity (r=−0.536), coupling DNA oxidative damage with antioxidant enzyme depletion. [3] In AD patients with concurrent physical frailty, urine 8-OHdG is elevated alongside inflammatory markers. [4]

安全提醒（方法學） Early ELISA studies may overestimate plasma/serum 8-OHdG by 3–10× due to nuclease-induced ex-vivo oxidation during sample processing. Urinary 8-OHdG is more stable; HPLC-ECD on urine samples is the most reproducible. Absolute values across studies are not directly comparable. Absolute values for serum 8-OHdG in the Cao 2020 cohort [2] were not reported — comparative analysis limited to direction and correlation.
StudyCompartmentN (cases/ctrl)Key FindingGRADE
Abe 2002 [1]CSF 8-OHG18 AD / 8 ctrl500±213 pM vs 97±32 pM; ~5× ↑ (P<0.001); rs=0.67 with MMSELow ⊕⊕⊝⊝
Cao 2020 [2]Serum 8-OHdG131 AD / 121 MCI / 100 ctrlControl < MCI < AD gradient (P<0.05); negative correlation with MoCA & MMSELow ⊕⊕⊝⊝
Zengi 2012 [3]Urine 8-OHdGAD vs ctrlUrine 8-OHdG ↑ in AD; r=−0.536 with PON1Low ⊕⊕⊝⊝
Namioka 2017 [4]Urine 8-OHdGAD frail vs non-frailAssociated with physical frailty + inflammation in ADLow ⊕⊕⊝⊝
Dai 2024 [6]Plasma 8-OHdGn=1,312 community elderlyOR 1.007 per unit ↑ for MCR (P=0.003)Low ⊕⊕⊝⊝
重點 • CSF 8-OHG（RNA 氧化）在 AD 有完整群體分離（5 倍升高），但為單一小型研究，且測量的是 RNA 損傷而非 DNA。
• 血清 8-OHdG 呈 Control < MCI < AD 劑量效應梯度，是 MCI 早期識別的潛在指標。
• 尿液 8-OHdG 與抗氧化酶（PON1）及身體衰弱呈負相關，代表氧化壓力的系統性影響。
• 所有 AD 相關研究均為觀察性，缺乏正式 ROC/AUC 分析，GRADE 均為 Low。
結論 8-OHdG/8-OHG elevations are consistently observed in AD across three compartments (CSF, serum, urine) and in prodromal MCI — supporting an early role of nucleic acid oxidative damage in AD pathogenesis. Formal diagnostic accuracy data (AUC, sensitivity, specificity) are lacking across all compartments; this is a key evidence gap.
03

8-OHdG in Parkinson's Disease PD 之 8-OHdG

證據顯示 CSF 8-OHdG (true DNA oxidation) is significantly elevated in PD versus controls, with a strong positive correlation with disease duration (rs=0.87, P<0.001) in 20 PD patients — a pattern opposite to AD where CSF 8-OHG inversely correlates with duration. [7] CoQ-10 oxidation (%CoQ-10: 80.3±17.9% vs 68.2±20.4%, P<0.05) parallels 8-OHdG elevation, linking mitochondrial dysfunction to DNA oxidative damage in early PD.

A larger study (n=44 PD, n=32 controls) confirmed CSF 8-OHdG and 8-OHG elevations in PD (P=0.02 and 0.04 respectively), [8] with an important disease-stage dissociation: 8-OHdG is elevated specifically in PD without dementia (P=0.05), suggesting it marks early neurodegeneration; conversely, CSF 8-OHG is lower in PD with dementia versus controls (P=0.04), possibly reflecting exhausted RNA production from dying neurons.

Urinary 8-OHdG (classic reference): In 72 PD patients, urinary 8-OHdG increases monotonically with H&Y stage and is not influenced by levodopa dose, [9] making it a candidate stage-tracking biomarker independent of treatment.

證據顯示 — 2026 Meta-analysis The most recent and largest meta-analysis (Msigwa et al. 2026; 54 studies, n=722 PD + 3,277 controls) found blood-based 8-OHdG moderately elevated in PD: Hedges' g=0.78 (95% CI 0.18–1.39; P=0.011). [10] Critically, F2-isoprostanes (lipid peroxidation) were NOT significantly elevated in PD (g=0.47; 95% CI −0.43–1.38; P=NS), suggesting DNA oxidative damage is more prominent than lipid peroxidation in PD — distinct from the combined elevation seen in type 2 diabetes. High heterogeneity (I²>90%) reflects assay and compartment variability.
Study / SourceCompartmentNKey FindingGRADE
Isobe 2010 [7]CSF 8-OHdG20 PD / 20 ctrl↑ in PD (P<0.0001); rs=0.87 with disease durationModerate ⊕⊕⊕⊝
Gmitterová 2018 [8]CSF 8-OHdG & 8-OHG44 PD / 32 ctrl8-OHdG ↑ early PD; 8-OHG ↓ in PD with dementiaModerate ⊕⊕⊕⊝
Sato 2005 [9] (classic)Urine 8-OHdG72 PDIncreases with H&Y stage; levodopa-independentVery Low ⊕⊝⊝⊝
Msigwa 2026 [10] (SR/meta)Blood 8-OHdG722 PD / 3,277 ctrlg=0.78 (95% CI 0.18–1.39; P=0.011); I²>90%Low ⊕⊕⊝⊝
重點 • CSF 8-OHdG 在 PD 早期（無失智）升高，且與病程正相關——與 AD 的「早升晚降」模式相反，提示 PD 中氧化 DNA 損傷隨疾病進展持續累積。
• 尿液 8-OHdG 追蹤 H&Y 分期，且不受左旋多巴影響，具備獨立 staging biomarker 的潛力（但僅一篇 classic reference，GRADE Very Low）。
• 2026 meta-analysis 確認血漿 8-OHdG 在 PD 有中度升高（g=0.78），但異質性極高；F2-isoprostanes 不顯著——氧化損傷類型可能有疾病特異性。
結論 8-OHdG in PD is supported at Moderate certainty in CSF (two consistent studies, strong correlation with disease duration) and Low certainty in blood (meta-analysis significant but I²>90%). A stage-stratified study combining CSF, urine, and blood compartments with standardised assays would substantially upgrade evidence quality.
04

Aging & 8-OHdG Confounders 老化與混淆因素

Oxidative DNA damage accumulates with normal aging. In a cross-sectional cohort (n=198, age 20–89), urinary 8-oxoG and 8-OHdG correlated positively with chronological age; a composite panel including 8-oxoG and dityrosine (DTyr) predicted accelerated biological aging with >92% accuracy, positioning urinary 8-OHdG as a biomarker of biological age rather than only disease. [11]

Critically, CSF 8-OHG did not correlate with age in healthy controls (rs=−0.25, P=0.35), [1] suggesting that CSF may be more disease-specific and less confounded by normal aging than urine or plasma. This has direct methodological implications: studies using peripheral 8-OHdG must age-match controls rigorously; CSF-based studies have an inherent advantage.

重點 • 尿液 8-OHdG 隨年齡上升（正常老化效應），研究設計必須嚴格年齡配對。
• CSF 8-OHG/8-OHdG 與年齡無顯著相關（對照組），CSF 測量具較高的疾病特異性。
• 其他混淆因素：腎功能（影響尿液排泄）、抽菸、運動量、飲食、慢性發炎、T2DM（Msigwa meta-analysis 發現 T2DM 8-OHdG 升高更顯著：g=2.64 vs PD g=0.78）。
結論 Age is a major confounder for peripheral 8-OHdG. Future diagnostic studies should report age-adjusted estimates or use CSF measurements to minimize this confound. The Mukli 2022 [11] urinary panel may serve as a biological aging index, complementary to disease-specific biomarkers.
05

Mitochondrial DNA in AD & MCI AD / MCI 之 mtDNA

證據顯示 — Blood mtDNA CN & Cognition In the largest meta-analysis to date (TOPMed, 9 diverse community cohorts, up to 19,152 participants), higher whole-blood mtDNA copy number (CN) was cross-sectionally associated with better general cognitive function (β=0.04; 95% CI 0.02–0.06). [12] Prospective analyses (5–20 years follow-up) showed similar but attenuated associations. Mendelian randomisation did not support a causal relationship, indicating mtDNA CN is a correlate, not a driver, of cognition. Blood mtDNA CN declines with normal aging across all cohorts.

Blood mtDNA CN also associates with brain tissue changes in AD: post-mortem brain tissue from AD patients shows up to 14% lower mtDNA CN compared to controls, with associations to tau and TDP-43 pathology. [12] A longitudinal study (8 years, n=75) found that individuals converting to AD show decreased D-loop methylation and increased mtDNA CN over time, while healthy controls show progressive D-loop methylation increase (possibly a protective mechanism). [16] These seemingly contradictory findings (lower mtDNA CN in brain tissue but higher peripheral CN in converters) may reflect different stages: early compensatory mtDNA replication followed by depletion with disease progression. In MCI, higher plasma ccf-mtDNA is observed specifically in APOE-ε4 carriers (P=0.05), suggesting gene–environment interaction in mitochondrial stress. [17]

安全提醒（矛盾發現） CSF cf-mtDNA in AD shows contradictory findings across studies. Podlesniy 2016 [13] and Podlesniy 2020 [15] report decreased CSF cf-mtDNA in slow-progressive AD (spAD); Cervera-Carles 2017 [14] found increased CSF mtDNA in AD (AUC=0.715) but with high overlap. Possible explanations: (1) disease stage — early depletion vs late release from dying neurons; (2) spAD vs rpAD heterogeneity; (3) sample preparation differences.

CSF cf-mtDNA in slow-progressive AD (spAD): cf-mtDNA was 44% lower than controls in spAD (69% lower in biomarker-selected cohort); in rapid-progressive AD (rpAD), no significant difference was found. [15] The cf-mtDNA / p-tau ratio achieved sensitivity 93%, specificity 94% for spAD in a biomarker-selected cohort (n=95 total, 30 spAD, 16 rpAD, 49 controls) — one of the highest reported diagnostic performances for a novel CSF biomarker in AD. This ratio is not yet validated in external cohorts.

Salivary mtDNA (non-invasive): In cognitively normal older adults, salivary mtDNA CN was positively correlated with cortical amyloid-β (Aβ) burden by PET and with plasma pTau-181, and negatively correlated with cognitive scores, [18] suggesting that salivary mtDNA tracks early Aβ pathology. It was not associated with NfL or GFAP — supporting a role complementary to, rather than redundant with, established biomarkers.

StudyMarker / CompartmentNKey FindingGRADE
Zhang 2023 [12] (meta)Blood mtDNA CNUp to 19,152β=0.04 (95% CI 0.02–0.06) with cognition; MR non-causalModerate ⊕⊕⊕⊝
Rizzo 2026 [16]Blood mtDNA D-loop / CN75 (8-yr longitudinal)AD converters: D-loop methylation↓ + CN↑; opposite in HCLow ⊕⊕⊝⊝
Podlesniy 2020 [15]CSF cf-mtDNA95 (30 spAD)spAD: −44% (−69% biomarker-selected); mtDNA/p-tau ratio: Sn 93%, Sp 94%Very Low ⊕⊝⊝⊝
Cervera-Carles 2017 [14]CSF cf-mtDNA124 AD / 140 ctrlAD: mtDNA ↑ (P sig); AUC=0.715; high overlapVery Low ⊕⊝⊝⊝
Choi 2024 [17]Plasma ccf-mtDNA332 MCI ± rMDDPlasma ccf-mtDNA ↑ in APOE-ε4 carriers (MCI only; P=0.05)Very Low ⊕⊝⊝⊝
Cantero 2025 [18]Salivary mtDNA CNCNA older adultsCorrelated with Aβ PET & pTau-181; not with NfL/GFAPLow ⊕⊕⊝⊝
Huang 2023 [30]Blood mtDNA indicatorsAD patientsmtDNA indicators correlate with cytokine profiles in AD immune dysregulationVery Low ⊕⊝⊝⊝
重點 • 血液 mtDNA CN 與認知功能正相關（Moderate 等級），但 MR 分析不支持因果性——mtDNA CN 是認知健康的「指標」而非「驅動因素」。
• CSF cf-mtDNA 在 AD 方向矛盾（spAD 偏低 vs Cervera-Carles 偏高），GRADE Very Low；cf-mtDNA/p-tau 比值（Sn 93%/Sp 94%）令人振奮但需外部驗證。
• 唾液 mtDNA 在認知正常者與 Aβ 及 pTau-181 相關，為非侵入性早期篩檢的新方向。
• APOE-ε4 基因型調節 MCI 的 ccf-mtDNA 水準，提示個別化生物標誌物解讀的必要性。
結論 The most robust mtDNA finding in AD is the blood mtDNA CN–cognition association (n=19,152; Moderate certainty); the absence of MR causality does not diminish its value as a biomarker. CSF cf-mtDNA data are inconsistent (Very Low certainty) and await standardized large-cohort replication. The cf-mtDNA/p-tau ratio and salivary mtDNA are promising but exploratory.
06

Mitochondrial DNA in PD & iRBD PD & iRBD 之 mtDNA

證據顯示 — Blood mtDNA CN in PD In the largest investigation of mtDNA CN in PD to date (n=363 peripheral blood + 151 substantia nigra pars compacta + 120 frontal cortex), PD patients show significant reduction in blood mtDNA CN and substantia nigra mtDNA CN versus matched controls; frontal cortex CN was unchanged. [19] The nigral reduction is disease-specific (not general brain degeneration) and is reflected in peripheral blood, validating blood mtDNA CN as a viable PD biomarker.

CSF cf-mtDNA in idiopathic PD (iPD): CSF cf-mtDNA is consistently reduced in iPD across three independent studies. [20][21][22] A key mechanistic insight: LRRK2-associated PD shows elevated CSF mtDNA, opposite to iPD, [20] and iPD is characterised by high cf-mtDNA deletion ratios whereas LRRK2-PD has no deletions. [21] This bimodal pattern suggests fundamentally different mitochondrial pathologies: copy-number depletion with structural damage in iPD versus increased mtDNA release without deletion in LRRK2-PD. A recent study (n=44 PD, n=43 controls) confirmed lower CSF mt64-ND1 and mt96-ND5 in PD (P=0.002, P=0.001) and identified body composition and serum albumin as key determinants — nutritional status is a critical confound for CSF cf-mtDNA in PD. [22]

安全提醒 — iRBD (Prodromal Lewy Body Disease) In a prospective study of 71 participants (17 non-converters, 34 converters to PD/DLB, 20 controls), iRBD patients — regardless of conversion status — had more cf-mtDNA deletions in CSF and reduced cf-mtDNA copies in CD9-positive extracellular vesicles (EVs). [23] Paradoxically, serum cf-mtDNA was INCREASED in iRBD converters (opposite to CSF direction). This blood/CSF discordance likely reflects different release mechanisms: EV-mediated release to CSF (reduced capacity due to mitochondrial dysfunction) versus passive leak across damaged membranes into blood. mtDNA dysfunction precedes motor symptom onset — it is not a consequence of disease conversion.

Post-mortem ventricular CSF confirmed reduced vCSF-cfmtDNA specifically in PD, not in other neurodegenerative diseases (AD, DLB, MSA) — suggesting PD-specific mitochondrial biology, though higher levels correlated with more severe clinical presentations (complex relationship, not simple depletion). [24] The Risi 2025 SR [26] (Eur J Neurol) confirmed mtDNA alterations are most consistently found in PD across studies, with blood intracellular and cf-mtDNA studies showing poor reproducibility due to lack of standardisation.

StudyMarkerSubtypeKey FindingGRADE
Pyle 2016 [19]Blood mtDNA CNiPD (n=363)↓ in blood and SN; frontal cortex sparedModerate ⊕⊕⊕⊝
Podlesniy 2016b [20]CSF mtDNAiPD ↓ vs LRRK2 ↑LRRK2-PD: high CSF mtDNA; iPD: low — opposite patternsModerate ⊕⊕⊕⊝
Puigròs 2022 [21]CSF cf-mtDNA deletionsiPD (n=validated)High deletion ratio in iPD; no deletions in LRRK2-PDModerate ⊕⊕⊕⊝
Mizutani 2025 [22]CSF mt-ND1/ND5iPD (n=44)↓ in PD (P=0.001); body composition & albumin = key confoundsModerate ⊕⊕⊕⊝
Puigròs 2024 [23]CSF cf-mtDNA deletions & serumiRBD converters / non-convertersCSF deletions ↑ + CD9-EV CN ↓ in all iRBD; serum CN ↑ in convertersLow ⊕⊕⊝⊝
Lowes 2020 [24]Post-mortem vCSF cf-mtDNAPD-specific ↓Reduced only in PD among NDDs; higher = more severe presentationLow ⊕⊕⊝⊝
重點 • iPD 的 CSF cf-mtDNA 在多個獨立研究中一致偏低，且伴有高缺失率（deletion ratio）——提示 mtDNA 結構性損傷，非只是複製數減少。
• LRRK2-PD 與 iPD 的 mtDNA 模式截然相反，可能用於鑑別診斷（需前瞻性驗證）。
• iRBD 患者 CSF cf-mtDNA 損傷已先於 PD 轉換出現——粒線體 DNA 功能障礙是 Lewy 體病的前驅性事件。
• 血清與 CSF cf-mtDNA 方向相反（iRBD 中），需同時測量雙compartment 才能正確解讀。
• 體組成和白蛋白是重要混淆因素（Mizutani 2025），臨床研究需控制。
結論 CSF cf-mtDNA reduction in iPD has Moderate certainty (3+ consistent studies; disease-specific pattern distinct from LRRK2-PD). The iRBD findings are compelling (Low certainty, needs replication) and position cf-mtDNA deletions as a prodromal biomarker for Lewy body disease.
07

Combination Biomarker Potential 組合標誌物潛力

No formal head-to-head study has compared 8-OHdG + mtDNA against established biomarkers (p-tau217, NfL, α-synuclein). Current indirect evidence suggests complementarity rather than redundancy:

CombinationStudyPerformanceNote
CSF cf-mtDNA + p-tauPodlesniy 2020 [15]Sn 93%, Sp 94% for spAD (biomarker-selected cohort)Exploratory; n=30 spAD; needs external validation
Salivary mtDNA + pTau-181Cantero 2025 [18]Correlated with Aβ PET; NOT with NfL/GFAPDistinct biological signal from standard biomarkers
CSF 8-OHdG + %CoQ-10Isobe 2010 [7]rs=0.56 (both elevated in early PD)Complementary mitochondrial dysfunction signals
Blood mtDNA CN (population)Zhang 2023 [12]β=0.04 for cognition; MR non-causalCommunity risk stratification, not diagnosis
Di Lorenzo pilot (CSF SOD2 + cf-mtDNA)Di Lorenzo 2025 [27]Male-specific SOD2 ↑ in MCI; ADd plasma DNase I & MMP-2 ↑Small pilot (n=20/group); sex-stratified effects
重點 • 8-OHdG 反映「下游氧化損傷」；mtDNA CN 反映「上游粒線體生合成能力」——兩者在機制上互補。
• cf-mtDNA/p-tau ratio（spAD）和唾液 mtDNA 的初步數據令人振奮，但樣本太小，不足以改變臨床實踐。
• 最重要的研究缺口：缺乏 8-OHdG 和 mtDNA 標誌物的正式 AUC 分析以及與 p-tau217、NfL、Aβ42/40 的頭對頭比較。
結論 The combination of oxidative (8-OHdG) and mitochondrial (cf-mtDNA) biomarkers represents a mechanistically distinct and potentially complementary layer to the Alzheimer's AT(N) and PD biomarker framework. Prospective studies with standardised assays, biomarker-selected cohorts, and comparison with established biomarkers are needed to establish clinical utility.
08

Summary & Research Gaps 結語 & 研究缺口

Key ClaimStrongest EvidenceGRADEEvidence Gap
CSF 8-OHG ↑ 5× in AD (early stage)Abe 2002 [1]: 500 vs 97 pMLow ⊕⊕⊝⊝Single small study; no independent replication with UPLC-MS/MS
Serum 8-OHdG gradient: Ctrl<MCI<ADCao 2020 [2]: n=352, P<0.05Low ⊕⊕⊝⊝No AUC; no standardised assay; absolute values missing
CSF 8-OHdG ↑ in PD; stage correlatesIsobe 2010 [7]: rs=0.87; Gmitterová 2018 [8]Moderate ⊕⊕⊕⊝AUC vs established markers; staging study with ddPCR assay
Blood 8-OHdG ↑ in PD (meta)Msigwa 2026 [10]: g=0.78 (0.18–1.39)Low ⊕⊕⊝⊝I²>90%; standardised assay protocol urgently needed
Blood mtDNA CN ↑ → better cognitionZhang 2023 [12]: β=0.04 (0.02–0.06), n=19,152Moderate ⊕⊕⊕⊝MR non-causal; no causal mechanism established
CSF cf-mtDNA ↓ in spADPodlesniy 2020 [15]: −44%; cf-mtDNA/p-tau Sn93%/Sp94%Very Low ⊕⊝⊝⊝Contradicted by Cervera-Carles 2017; needs large validation
CSF cf-mtDNA ↓ in iPDPodlesniy 2016b [20], Puigròs 2022 [21], Mizutani 2025 [22]Moderate ⊕⊕⊕⊝Nutritional confounders; LRRK2 differentiation needs prospective validation
CSF cf-mtDNA deletions in iRBD precede PDPuigròs 2024 [23]: n=71, prospectiveLow ⊕⊕⊝⊝Single study; needs larger prodromal cohort
Salivary mtDNA tracks Aβ + pTauCantero 2025 [18]: correlates with Aβ PETLow ⊕⊕⊝⊝Cross-sectional; no longitudinal conversion data
8 Take-home Messages 1. 8-OHG（RNA 氧化）和 8-OHdG（DNA 氧化）在 AD CSF 中均升高，但兩者生化意涵不同，研究中必須明確區分。
2. 血清 8-OHdG 在 AD/MCI 中呈劑量效應梯度，是早期 MCI 識別的潛在標誌物，但缺乏正式 AUC 分析。
3. CSF 8-OHdG 在 PD 持續升高且與病程強正相關（rs=0.87），與 AD 早升晚降的模式相反。
4. 血液 mtDNA CN 與認知功能正相關（n=19,152，Moderate），但 MR 非因果——是反映指標而非驅動機制。
5. CSF cf-mtDNA 在 AD 方向矛盾（spAD↓ vs Cervera-Carles↑），GRADE Very Low；cf-mtDNA/p-tau 比值（Sn93%/Sp94%）需外部驗證。
6. CSF cf-mtDNA ↓ 於 iPD，Moderate 等級；LRRK2-PD 卻↑——可作兩型 PD 鑑別潛在工具。
7. iRBD 的 CSF cf-mtDNA 缺失已在轉換前出現，定位粒線體 DNA 損傷為 Lewy 體病前驅機制。
8. 8-OHdG + cf-mtDNA 與建立標誌物（p-tau217、NfL）的頭對頭比較研究完全缺乏，是最重要的研究缺口。
REF

References 參考文獻 (30 篇，PMID 驗證)

  1. Abe T, Tohgi H, Isobe C, Murata T, Sato C. Remarkable increase in the concentration of 8-hydroxyguanosine in cerebrospinal fluid from patients with Alzheimer's disease. J Neurosci Res. 2002;70(3):447–50. PMID 12391605
  2. Cao X, Chen P. Changes in Serum Amyloid A (SAA) and 8-OHdG in Patients with Senile Early Cognitive Impairment. Med Sci Monit. 2020;26:e919586. PMID 32006457
  3. Zengi O, Karakas A, Ergun U, Senes M, Inan L, Yucel D. Urinary 8-hydroxy-2'‑deoxyguanosine level and plasma paraoxonase 1 activity with Alzheimer's disease. Clin Chem Lab Med. 2012;50(3):529–34. PMID 22098435
  4. Namioka N, Hanyu H, Hirose D, et al. Oxidative stress and inflammation are associated with physical frailty in patients with Alzheimer's disease. Geriatr Gerontol Int. 2017;17(6):913–8. PMID 27296166Local Drive
  5. Peña-Bautista C, Tirle T, López-Nogueroles M, et al. Oxidative Damage of DNA as Early Marker of Alzheimer's Disease. Int J Mol Sci. 2019;20(24):6136. PMID 31817451
  6. Dai Q, Ma Y, Liu C, et al. Association of 8-hydroxy-2'‑deoxyguanosine with motoric cognitive risk in elderly Chinese people. BMC Geriatr. 2024;24(1):331. PMID 38605326
  7. Isobe C, Abe T, Terayama Y. Levels of reduced and oxidized coenzyme Q-10 and 8-OHdG in the CSF of patients with living Parkinson's disease. Neurosci Lett. 2010;469(1):159–63. PMID 19944739
  8. Gmitterová K, Gawinecka J, Heinemann U, Valkovič P, Zerr I. DNA versus RNA oxidation in Parkinson's disease: Which is more important? Neurosci Lett. 2018;662:22–28. PMID 28963060
  9. Sato S, Mizuno Y, Hattori N. Urinary 8-hydroxydeoxyguanosine levels as a biomarker for progression of Parkinson disease. Neurology. 2005;64(6):1081–3. PMID 15781836
  10. Msigwa SS, Mkwambe MC, Yu QY, et al. Comparative evaluation of oxidative stress biomarkers F2-isoprostanes and 8-OHdG in Parkinson's disease and Type 2 Diabetes Mellitus: a systematic review and meta-analysis of human studies. Ann Med. 2026;58(1):2654251. PMID 42003321
  11. Mukli P, Wu DH, Csipo T, et al. Urinary Biomarkers of Oxidative Stress in Aging: Implications for Prediction of Accelerated Biological Age in Prospective Cohort Studies. Oxid Med Cell Longev. 2022:6110226. PMID 35571254
  12. Zhang Y, Liu X, Wiggins KL, et al. Association of Mitochondrial DNA Copy Number With Brain MRI Markers and Cognitive Function: A Meta-analysis of Community-Based Cohorts. Neurology. 2023;100(18):e1930–43. PMID 36927883Local Drive
  13. Podlesniy P, Llorens F, Golanska E, et al. Mitochondrial DNA differentiates Alzheimer's disease from Creutzfeldt-Jakob disease. Alzheimers Dement. 2016;12(5):546–55. PMID 26806388
  14. Cervera-Carles L, Alcolea D, Estanga A, et al. Cerebrospinal fluid mitochondrial DNA in the Alzheimer's disease continuum. Neurobiol Aging. 2017;53:192.e1–e4. PMID 28089353Local Drive
  15. Podlesniy P, Llorens F, Puigròs M, et al. Cerebrospinal Fluid Mitochondrial DNA in Rapid and Slow Progressive Forms of Alzheimer's Disease. Int J Mol Sci. 2020;21(17):6298. PMID 32878083Local Drive
  16. Rizzo B, Rossi M, Ferrari RR, et al. Longitudinal Analysis of Mitochondrial D-Loop Methylation and Copy Number in Peripheral Blood: Epigenetic Signatures of Alzheimer's Disease Progression and Aging. Int J Mol Sci. 2026;27(3):1477. PMID 41683898
  17. Choi J, Beroncal EL, Chernega T, et al. Exploring mitochondrial blood-based and genetic markers in older adults with mild cognitive impairment and remitted major depressive disorder. Transl Psychiatry. 2024;14(1):457. PMID 39468012
  18. Cantero JL, Atienza M, Podlesniy P, Puigròs M, Trullas R. Salivary mitochondrial DNA is associated with biomarkers of Alzheimer's disease in cognitively normal older adults. Transl Psychiatry. 2025;15(1):355. PMID 41052979
  19. Pyle A, Anugrha H, Kurzawa-Akanbi M, Yarnall A, Burn D, Hudson G. Reduced mitochondrial DNA copy number is a biomarker of Parkinson's disease. Neurobiol Aging. 2016;38:216.e7–e10. PMID 26639155
  20. Podlesniy P, Vilas D, Taylor P, Shaw LM, Tolosa E, Trullas R. Mitochondrial DNA in CSF distinguishes LRRK2 from idiopathic Parkinson's disease. Neurobiol Dis. 2016;94:10–17. PMID 27260835
  21. Puigròs M, Calderon A, Pérez-Soriano A, et al. Cell-free mitochondrial DNA deletions in idiopathic, but not LRRK2, Parkinson's disease. Neurobiol Dis. 2022;174:105885. PMID 36208866
  22. Mizutani Y, Nakai T, Maeda Y, et al. CSF Mitochondrial DNA: Biomarker of Body Composition and Energy Metabolism in Parkinson's Disease. Ann Clin Transl Neurol. 2025;12(12):2410–21. PMID 40891077
  23. Puigròs M, Calderon A, Martín-Ruiz D, et al. Mitochondrial DNA deletions in the cerebrospinal fluid of patients with idiopathic REM sleep behaviour disorder. EBioMedicine. 2024;102:105065. PMID 38502973
  24. Lowes H, Kurzawa-Akanbi M, Pyle A, Hudson G. Post-mortem ventricular cerebrospinal fluid cell-free-mtDNA in neurodegenerative disease. Sci Rep. 2020;10(1):15253. PMID 32943697
  25. Gambardella S, Limanaqi F, Ferese R, et al. ccf-mtDNA as a Potential Link Between the Brain and Immune System in Neuro-Immunological Disorders. Front Immunol. 2019;10:1064. PMID 31143191Local Drive
  26. Risi B, Imarisio A, Cuconato G, et al. Mitochondrial DNA (mtDNA) as fluid biomarker in neurodegenerative disorders: A systematic review. Eur J Neurol. 2025;32(1):e70014. PMID 39831374Local Drive
  27. Di Lorenzo R, Zecca C, Chimienti G, et al. Reliable New Biomarkers of Mitochondrial Oxidative Stress and Neuroinflammation in CSF and Plasma from Alzheimer's Disease Patients: A Pilot Study. Int J Mol Sci. 2025;26(16):7792. PMID 40869114
  28. Cioffi F, Adam RHI, Bansal R, Bhatt D. A Review of Oxidative Stress Products and Related Genes in Early Alzheimer's Disease. J Alzheimers Dis. 2021;83(3):977–1000. PMID 34420962
  29. Stoccoro A, Smith AR, Baldacci F, et al. Mitochondrial D-Loop Region Methylation and Copy Number in Peripheral Blood DNA of Parkinson's Disease Patients. Genes (Basel). 2021;12(5):720. PMID 34065874
  30. Huang J, Song Z, Wei B, et al. Immunological evaluation of patients with Alzheimer's disease based on mitogen-stimulated cytokine productions and mitochondrial DNA indicators. BMC Psychiatry. 2023;23(1):160. PMID 36890488Local Drive