BASE問題與穩定度評估The Problem & Stability Assessment
感染除了要控制，也可能破壞結構、造成不穩定。第一個問題是：這個脊椎穩不穩？Infection must be controlled, but it can also destroy structure and cause instability. The first question: is this spine stable?
脊椎感染：現況與處置流程Spinal infection: state of the art and management algorithm
感染破壞椎間盤與終板，可能導致不穩定；治療須兼顧感染控制與穩定。Disc–endplate destruction can lead to instability; treatment must address both infection control and stability.
敘述性文獻回顧，附診療流程圖
Narrative review with a management algorithm
- 佔肌肉骨骼感染 2–7%；從症狀到確診常延遲 2–6 個月。S. aureus 佔 30–80%，約三分之一找不到病原。
- 病理機轉：血行播散 → endplate erosion、溶骨性病灶、壓迫性骨折 → instability、deformity、神經壓迫。
- MRI 為影像 gold standard（sensitivity 96%、specificity 94%）；建議常規取組織做病理與培養。
- 手術適應症：神經學缺損或 sepsis、骨破壞造成的 instability、嚴重 kyphosis、保守治療失敗。
- 2–7% of musculoskeletal infections; diagnosis is often delayed 2–6 months. S. aureus causes 30–80%; about one-third remain without an identified organism.
- Pathway: haematogenous seeding → endplate erosion, lytic lesions, compression fracture → instability, deformity, neural compromise.
- MRI is the imaging gold standard (sensitivity 96%, specificity 94%); routine tissue sampling for histology and culture is recommended.
- Surgical indications: neurological deficit or sepsis, instability from bone destruction, severe kyphosis, failed conservative care.
解讀限制Limitations
- 敘述性回顧，沒有系統性搜尋或證據分級；流程圖未經驗證。
- 2013 年發表，後路與微創手術已有較新證據。
- Narrative review without systematic search or evidence grading; the algorithm is unvalidated.
- Published 2013; newer evidence exists on posterior-only and minimally invasive surgery.
Duarte RM, Vaccaro AR. Eur Spine J. 2013;22:2787–2799. doi:10.1007/s00586-013-2850-1
mSISS：脊椎感染的不穩定評分mSISS: an instability score for spondylodiscitis
四項 CT／臨床參數分三級；專家間信度高，但尚未以臨床結果驗證。Four CT/clinical parameters, three categories; high expert reliability, but not yet validated against outcomes.
AO Spine 多中心 reliability 研究；15 位專家評估 10 例，間隔一個月評兩輪
AO Spine multicentre reliability study; 15 experts rated 10 cases twice, one month apart
Location0 / 3
Junctional（Occ–C2、C7–T2、T11–L1、L5–S1）3 分
Junctional (Occ–C2, C7–T2, T11–L1, L5–S1) scores 3
Bone lesion0 / 3
節段受累 >50% 為 3 分（以整個節段計，不限單一椎體）
>50% segmental involvement scores 3 (whole segment, not one vertebral body)
Alignment0 / 2 / 3
新發 kyphosis／scoliosis 2 分；subluxation／translation 3 分
De novo kyphosis/scoliosis 2; subluxation/translation 3
Mechanical pain0 / 1
與負重相關的疼痛 1 分
Load-related pain scores 1
- 總分 0–2 stable、3–4 potentially unstable、5–10 unstable。
- Interrater ICC 0.87／0.89（兩輪）；intrarater ICC 0.90。Alignment 一致性最低。
- 依評分規則：junctional 合併 >50% 受累即 6 分（unstable），即使排列正常。這是「排列正常不代表穩定」的依據。
- Total 0–2 stable, 3–4 potentially unstable, 5–10 unstable.
- Interrater ICC 0.87 / 0.89 (two rounds); intrarater ICC 0.90. Alignment had the lowest agreement.
- By the scoring rules, junctional location plus >50% involvement already scores 6 (unstable) even with normal alignment—the basis for “preserved alignment does not exclude instability.”
解讀限制Limitations
- 只有 10 例、15 位受過訓練的專家；未評估臨床結果或預測效度。
- 只適用 pyogenic，不適用 TB；作者強調不能單獨決定是否開刀。
- Only 10 cases and 15 trained experts; no outcome or predictive validation.
- Pyogenic infection only (not TB); the authors stress it is not a stand-alone surgical indication.
Schömig F, Taheri N, El-Sharkawi M, et al. The Modified Spinal Instability Spondylodiscitis Score (mSISS). Global Spine J. 2026; OnlineFirst. doi:10.1177/21925682261455309
FIXATION固定：可不可以、要多長？Can we instrument, and how long?
活動性感染不是 instrumentation 的絕對禁忌；固定長度、要不要在感染椎體打 screw，都應個別化。Active infection is not an absolute contraindication to instrumentation; construct length and screw placement in infected vertebrae should be individualised.
血行性脊椎骨髓炎：清創加內固定的結果Haematogenous vertebral osteomyelitis: outcomes with instrumentation
清創加內固定，復發率與單純清創無顯著差異；抗生素療程宜 ≥6 週。Recurrence after debridement plus instrumentation did not differ significantly from debridement alone; antibiotics ≥6 weeks matter.
南韓 5 家醫院回溯性世代；153 位接受手術清創的 HVO 病人（單純清創 94、清創加內固定 59）
Retrospective cohort, 5 Korean hospitals; 153 surgically debrided HVO patients (debridement alone 94, plus instrumentation 59)
單純清創Debridement alone清創＋內固定+ Instrumentation
感染復發Recurrence4.8% (4/84)6.8% (4/59)P = .72
- Event-free survival 無差異（P = .41）；single-stage 與 two-stage 的復發率也無差異。
- 內固定組抗生素療程越長、復發越少：4–6 週 22.2%、6–8 週 9.1%、≥8 週 2.6%（P = .04）。
- No difference in event-free survival (P = .41); single- vs two-stage surgery did not differ in recurrence.
- In the instrumented group, longer antibiotics meant fewer recurrences: 4–6 wk 22.2%, 6–8 wk 9.1%, ≥8 wk 2.6% (P = .04).
解讀限制Limitations
- 「無內固定」組也做了手術清創，不是保守治療組。
- 只有 8 件復發，檢定力有限；P = .72 不代表等效。
- 回溯性研究；以 S. aureus 為主（MRSA 38.3%）。
- The non-instrumented group was also debrided surgically—not a non-operative group.
- Only 8 recurrences; P = .72 does not prove equivalence.
- Retrospective; mostly S. aureus (MRSA 38.3%).
Park KH, Cho OH, Lee YM, et al. Clin Infect Dis. 2015;60(9):1330–1338. doi:10.1093/cid/civ066
脊椎感染使用內固定的安全性Safety of instrumentation in spinal infection
清創後放內固定，觀察到的復發率低；評估不穩定要連清創後的狀態一起考慮。Low observed recurrence after instrumentation following debridement; assess instability including what debridement will create.
美國單中心回溯性病例系列，47 位清創後置入內固定，附文獻回顧
Single-centre US case series, 47 patients instrumented after debridement, plus literature review
- 感染復發 2/47，均為後方內固定；其中 1 位需移除內固定。
- 不穩定有兩個來源：術前椎體與椎間盤破壞，以及術中 radical resection。評估要同時看「現有」與「清創後」。
- 文獻回顧：689 位內固定病人中 12 例（1.74%）復發或因感染移除 implant。
- Recurrence 2/47, both with posterior instrumentation; one required hardware removal.
- Instability arises from preoperative vertebral/disc destruction and from radical intraoperative resection—assess both existing and post-debridement stability.
- Literature review: 12 of 689 instrumented patients (1.74%) had recurrence or implant removal for infection.
解讀限制Limitations
- 沒有無內固定的對照組。
- 約 58% 為頸椎、多為靜脈藥癮者，和典型胸腰椎感染不同。
- No non-instrumented control group.
- ~58% cervical and mostly IV drug users—unlike typical thoracolumbar disease.
Rayes M, Colen CB, Bahgat DA, et al. J Neurosurg Spine. 2010;12:647–659. doi:10.3171/2009.12.SPINE09428
在感染椎體置入 pedicle screwPedicle screws in the infected vertebra
CT 上確認 screw 軌道經過完好骨質時，可打在感染椎體，換來較短手術與固定範圍。When CT shows the screw path runs through preserved bone, screws can go into the infected vertebra—shorter surgery and constructs.
日本 9 家醫院回溯性世代；70 位微創後方固定（未清創）的胸腰椎感染病人
Retrospective cohort, 9 Japanese centres; 70 thoracolumbar cases with minimally invasive posterior fixation (no debridement)
打在感染椎體 (n=36)Screws in infected level (n=36)跳過 (n=34)Skipped (n=34)
手術時間Operative time146 min195 minP = .032
固定範圍Fixation range5.4 節5.4 vertebrae6.9 節6.9 vertebraeP = .0009
非計畫再手術Unplanned reoperation8/369/34P = .78
解讀限制Limitations
- 打 screw 組骨破壞較輕（selection bias），沒有做 matching。
- 再手術無顯著差異不代表等效；病人未清創，不適用大範圍骨破壞。
- The screw group had less bone destruction (selection bias); no matching.
- No significant difference ≠ equivalence; no debridement, so not applicable to extensive destruction.
Gamada H, Funayama T, Setojima Y, et al. BMC Musculoskelet Disord. 2024;25:458. doi:10.1186/s12891-024-07565-0
ICM 2025 國際共識：何時固定、固定多長ICM 2025 consensus: when to fix, and how long
固定長度沒有統一標準，應依骨破壞程度與骨質個別化。There is no universal construct length; individualise to bone destruction and bone quality.
系統性文獻回顧＋國際專家匿名投票；證據等級依 GRADE
Systematic review plus anonymous international expert voting; GRADE evidence levels
Q6.1 何時固定Q6.1 When to instrumentModerate · 97.3%
出現 instability、神經學缺損、內科治療失敗或有變形風險時建議 instrumentation，且不增加 reinfection 風險。
Recommended for instability, neurological deficit, failed medical therapy or risk of deformity, without increasing reinfection risk.
Q6.4 固定節數Q6.4 Fixation levelsLow · 100%
依椎體破壞程度與 osteoporosis 等因素個別化。Rationale 提到：破壞有限、骨質好時 monosegmental 可能足夠；嚴重破壞、變形或骨質差時考慮上下各 2 節。
Individualise by vertebral destruction and factors such as osteoporosis. The rationale suggests monosegmental may suffice for limited destruction with good bone; consider two levels above and below for severe destruction, deformity or poor bone.
解讀限制Limitations
- 手術相關題目證據多屬 Low 至 Moderate，以回溯性研究為主。
- 「上下各 2 節」出自 rationale，不是正式建議文字。
- Surgical questions rest mostly on low-to-moderate, retrospective evidence.
- “Two levels above and below” comes from the rationale, not the formal recommendation.
Rajasekaran S, Muthu S, Gurusamy G, Parvizi J; ICM-Spinal Infection Collaborators. Global Spine J. 2026;16(3S):27S–70S. doi:10.1177/21925682261454361
FUSION融合：一定要植骨嗎？Is formal grafting always needed?
穩定固定後，部分病人不植骨也能癒合；含抗生素骨移植可能有輔助角色。After stable fixation, some patients achieve union without grafting; antibiotic-loaded graft may have an adjunctive role.
不植骨的微創後方固定：感染椎體的骨癒合Bone union after posterior fixation without grafting
單節感染只做後方固定、不植骨，多數能骨癒合；但沒有與植骨直接比較。Most single-level cases united after posterior fixation alone—but there was no grafted comparison group.
日本 6 家醫院回溯性研究；40 位只做後方固定（未清創、未植骨）、追蹤 ≥12 個月，以 CT 判定
Retrospective, 6 Japanese centres; 40 patients with posterior fixation only (no debridement or graft), ≥12-month CT follow-up
單節椎間盤Single disc兩節椎間盤Two discs
骨癒合率Union rate88% (30/34)33% (2/6)P = .0095
- 整體骨癒合率 80%（32/40），術後平均 10.7 個月。
- 全體 40 人中 26 人（65%）最早在外側或前方 bridging callus 癒合；沒有一例從椎間盤中央開始。
- 無感染復發；3 位因 screw 鬆脫需延長固定。
- Overall union 80% (32/40) at a mean 10.7 months.
- In 26 of all 40 patients (65%), union began as lateral or anterior bridging callus; none began centrally in the disc.
- No recurrence; 3 needed extended fixation for screw loosening.
解讀限制Limitations
- 單臂研究，沒有植骨對照組，不能說不植骨等同或不劣於植骨。
- 失追蹤率 47%（35/75）；兩節感染只有 6 例。
- Single-arm: cannot claim non-inferiority to grafted fusion.
- 47% lost to follow-up (35/75); only 6 two-disc cases.
Gamada H, Funayama T, Nagasawa K, et al. BMC Musculoskelet Disord. 2024;25:860. doi:10.1186/s12891-024-07993-y
TIDF 合併含抗生素骨移植（AIBG）TIDF with antibiotic-impregnated bone graft (AIBG)
AIBG 與較短的術後靜脈抗生素療程、較快的早期 CRP 下降有關。AIBG was associated with shorter postoperative IV antibiotics and faster early CRP decline.
高雄長庚回溯性研究；AIBG 30 例 vs 歷史對照 28 例，皆經後路 TIDF
Retrospective, Kaohsiung Chang Gung; AIBG 30 vs historical controls 28, all posterior TIDF
AIBG (n=30)對照 (n=28)Control (n=28)
術後靜脈抗生素Postoperative IV antibiotics19.0 天19.0 days39.8 天39.8 daysP < .05
術後第 6 週 CRPCRP at week 65.7 mg/L13.0 mg/LP = .002
- 第 12 週 CRP 兩組已無差異。AIBG 組 26/30（86.7%）達 solid fusion；論文沒有對照組融合率。
- 抗生素依培養選擇；無培養結果時用 vancomycin 2 g + gentamicin 80 mg。
- CRP no longer differed at week 12. Solid fusion in 26/30 (86.7%) AIBG patients; no control fusion rate reported.
- Antibiotics chosen by culture; vancomycin 2 g + gentamicin 80 mg when no culture was available.
解讀限制Limitations
- 兩組分屬不同術者、不同年代（非同期對照）。
- 靜脈療程較短不等於總抗生素療程較短；未測量局部抗生素濃度。
- Different surgeons and eras (non-contemporaneous controls).
- Shorter IV therapy ≠ shorter total antibiotics; local drug levels were not measured.
Chang CC, Chang HK, Lu ML, Wegner A, Wu RW, Yin TC. Asian Spine J. 2025;19(1):38–45. doi:10.31616/asj.2024.0388
RECONSTRUCTION重建：何時需要、哪條路、什麼材料？When, which approach, which material?
前柱重建不是每例都需要；需要時，入路與材料要配合缺損與骨質。前柱重建不一定要走前路。Anterior reconstruction is not always needed; when it is, match approach and material to the defect and bone quality. Anterior column reconstruction does not require an anterior approach.
ICM 2025 國際共識：前柱重建與植骨材料ICM 2025 consensus: anterior support and graft choice
嚴重前柱破壞才需重建；高度喪失約 50% 是實務參考，不是絕對門檻。Reconstruct for major anterior destruction; ~50% height loss is a practical guide, not an absolute threshold.
Q6.2 前柱／360° 重建Q6.2 Anterior / 360° reconstructionModerate · 83.8%
嚴重前柱破壞、instability、kyphotic deformity、膿瘍形成或神經壓迫時有適應症；哪種術式較佳，證據互相矛盾。
Indicated for severe anterior destruction, instability, kyphotic deformity, abscess formation or neurological compromise; evidence on which technique is better is conflicting.
Q6.3 植骨或 cageQ6.3 Graft or cageLow · 100%
高度喪失 <50% 可考慮非結構性植骨；更嚴重時考慮結構性植骨或 titanium mesh cage。Osteoporosis 放 cage 有 subsidence 風險。
<50% height loss: consider non-structural graft; more severe: structural graft or titanium mesh cage. Cages risk subsidence in osteoporosis.
解讀限制Limitations
- 原文明言缺乏公認的缺損大小標準；50%、30° 出自 rationale，不是驗證過的切點。
- The paper states there is no accepted defect-size standard; 50% and 30° come from the rationale and are not validated cut-offs.
Rajasekaran S, et al. Global Spine J. 2026;16(3S):27S–70S. doi:10.1177/21925682261454361
前後合併 vs 單純後路Combined anterior–posterior vs posterior-only
經過選擇的病人，單純後路手術負擔較低；嚴重塌陷或變形者，作者以前後合併處理。In selected patients posterior-only meant less operative burden; the authors used combined AP for major collapse or deformity.
單一術者回溯性比較（Level IV）；79 例，單純後路 20 vs 前後合併 59
Single-surgeon retrospective comparison (Level IV); 79 patients, posterior-only 20 vs combined AP 59
單純後路Posterior-only前後合併Combined AP
手術時間Operative time303 min463 minP < .001
估計失血量Blood loss752 mL913 mLP = .007
融合率Fusion rate100%94.9%P = .310
Kyphosis 矯正Kyphosis correction5.3°12.8°P = .003
- 作者選擇前後合併的條件：高度喪失 >50%、胸腰段 kyphosis >25°，或腰椎 lordosis 喪失。
- 前後合併組術前較嚴重（高度喪失 33.9% vs 18.1%）。
- Authors chose combined AP for height loss >50%, thoracolumbar kyphosis >25°, or loss of lumbar lordosis.
- The AP group was more severe at baseline (height loss 33.9% vs 18.1%).
解讀限制Limitations
- 兩組基準嚴重度不同，比的是「被選擇的術式」，不能證明哪種入路較好。
- 分組條件是本研究自訂；排除神經缺損與免疫低下病人。
- Baseline severity differed—this compares selected strategies, not which approach is better.
- Selection criteria are study-specific; neurological deficit and immunocompromise were excluded.
Choi SH, Koo JW, Hur JM, Kang CN. Spine. 2020;45(19):E1239–E1248. doi:10.1097/BRS.0000000000003544
活動性感染以鈦網籠重建前柱Titanium mesh cages in active infection
鈦網籠加自體骨重建前柱可行，多數可經後路完成；但沒有對照組。Titanium mesh cage with autograft is feasible, mostly via a posterior-only route—but uncontrolled.
印度單中心回溯性病例系列；70 例活動性結核或化膿性感染，cage（內填自體骨）＋後路固定
Single-centre Indian case series; 70 active TB or pyogenic infections, cage with autograft plus posterior instrumentation
- 58/70（82.8%）經單純後路完成前柱清創與 cage 置入。
- 骨融合 60/64（93.7%），平均追蹤 25 個月；4 例 pseudarthrosis（1 例合併 subsidence）。
- 1 例持續流膿形成瘻管，移除內固定後癒合。
- 58/70 (82.8%) had anterior debridement and cage placement through a posterior-only route.
- Fusion 60/64 (93.7%) at mean 25 months; 4 pseudarthroses (one with subsidence).
- 1 persistent sinus healed after implant removal.
解讀限制Limitations
- 42/70 病理符合結核，不是純化膿性族群。
- 沒有 iliac crest 對照組；只有 10 例以 CT 評估融合。
- 42/70 had TB histology—not a pure pyogenic cohort.
- No iliac-crest comparator; CT confirmed fusion in only 10 patients.
Sundararaj GD, Amritanand R, Venkatesh K, Arockiaraj J. Asian Spine J. 2011;5(3):155–161. doi:10.4184/asj.2011.5.3.155
結構性異體骨 vs 鈦網籠Structural allograft vs titanium cage
兩者皆可行，融合率差異未達顯著，兩組都沒有復發。Both were feasible; fusion did not differ significantly, and neither group had recurrence.
韓國單中心回溯性比較；33 例，均經後路清創加 pedicle screw 固定
Single-centre Korean retrospective comparison; 33 patients, all posterior debridement plus pedicle screws
屍體異體骨 (n=15)Allograft (n=15)鈦網籠＋自體骨 (n=18)Cage + autograft (n=18)
融合Fusion14/1515/18P = .397
Subsidence >5 mmSubsidence >5 mm7/155/18P = .276
感染復發Recurrence0/150/18—
解讀限制Limitations
- 樣本小、非等效性設計；零事件不代表感染風險相同。
- 多為術後感染（S. epidermidis 為主）；cage 組仍需取 iliac 骨。
- Small, not an equivalence design; zero events ≠ equal risk.
- Mostly postoperative infections (S. epidermidis); the cage group still needed iliac harvest.
Kim HW, Ryu JI, Bak KH. J Korean Neurosurg Soc. 2011;50(4):348–356. doi:10.3340/jkns.2011.50.4.348